Pharmacokinetics

The Journal of Clinical Pharmacology2016, 00(0) 1–12C© 2016, The American College ofClinical PharmacologyDOI: 10.1002/jcph.756

Long-Term Effect of Rifampicin-BasedAnti-TB Regimen Coadministrationon the Pharmacokinetic Parametersof Efavirenz and 8-Hydroxy-Efavirenz inEthiopian Patients

Abiy Habtewold, PhD1,2,3, Eleni Aklillu, PhD4, Eyasu Makonnen, PhD3,Wondwossen Amogne,MD, PhD5,Getnet Yimer,MD, PhD3,Getachew Aderaye,MD, PhD5, Leif Bertilsson, PhD2, Joel S.Owen, PhD1,and Jurgen Burhenne, PhD6

Abstract

We compared the pharmacokinetic (PK) exposure parameters of efavirenz (EFV) and its major inactive metabolite, 8-hydroxy-efavirenz (8-OH-EFV), inan open-label, single-sequence, and parallel design of HIV-infected and tuberculosis (TB)-HIV-coinfected Ethiopian patients in the HIV-TB Pharmagenestudy with 20 and 33 patients, respectively. Both treatment groups underwent PK sampling following oral 600 mg EFV in week 16 of initiating EFV-based combination antiretroviral therapy.The TB-HIV-coinfected group repeated the PK sampling 8 weeks after stopping rifampin (RIF)–based anti-TBtreatment. Between-treatment group analysis indicated no significant effect of RIF-based anti-TB cotreatment on PK exposure parameters of EFV, norwas there a significant effect after controlling for sex or CYP2B6 genotype.However, RIF-based therapy in TB-HIV-coinfected patients had significantlyincreased 8-OH-EFV PK exposure measures and metabolic ratio relative to HIV-only patients,AUC0–24 greater by 79%.The effect was more prominentin women and CYP2B6*6 carriers in within-sex and CYP2B6 genotype comparisons. Within-subject comparisons for AUC0–24 and Cmax when “on”and “off” RIF-based anti-TB cotreatment showed geometric mean ratios (90% confidence intervals) of 100.5% (98.7%–102.3%) and 100.2% (98.1%–102.4%), respectively, for EFV and 98.6% (95.5%–101.7%–) and 97.6% (92.2%–103.0%), respectively, for 8-OH-EFV.We report no significant influenceof RIF-based anti-TB cotherapy on the EFV PK exposure measures. The study also calls for caution related to higher exposure to 8-OH-EFV duringsimultaneous coadministration of EFV and RIF-based anti-TB regimens, which may be associated with neurotoxicity, particularly in female patients andCYP2B6*6 carriers.

Keywords

efavirenz, 8-hydroxy-efavirenz, rifampicin, CYP2B6, TB-HIV, Ethiopian

Debate regarding the effect of rifampicin (RIF)–basedanti-TB cotreatment on efavirenz (EFV) plasma expo-sure and whether there is a need for EFV dose esca-lation to 800 mg/day in HIV patients being cotreatedfor TB with RIF-based anti-TB regimens remains un-resolved. Prior studies recommended an increased EFVdose of 800 mg per day when coadministered withRIF-based anti-TB treatment.1,2 In contrast, severalrecent studies including ours3–12 showed that RIF didnot significantly modify the exposure and subsequentefficacy of EFV when both were coadministered. Fur-thermore, the effect of a RIF-based anti-TB regimenon the main metabolite of EFV exposure is not wellinvestigated.

EFV is primarily metabolized by a polymorphicenzyme, CYP2B6, to 8-hydroxy-EFV (8-OH-EFV),a pharmacologically inactive metabolite but recentlyassociated with neurotoxicity, with minor metabolismby CYP3A4/5.13,14 Though minimal, EFV is also

metabolized to 7-hydroxy-EFV (7-OH-EFV) viaCYP2A6 metabolic pathway,14,15 and EFV also

1Department of Pharmaceutical Sciences, School of Pharmacy, UnionUniversity, Jackson, TN, USA2Division of Clinical Pharmacology, Department of Lab Medicine,Karolinska Institutet Hospital Huddinge, Stockholm, Sweden3Department of Pharmacology, School of Medicine,Addis Ababa Univer-sity, Addis Ababa, Ethiopia4Section of Pharmacogenetics, Department of Physiology & Pharmacol-ogy, Karolinska Institutet, Stockholm, Sweden5Department of Internal Medicine, School of Medicine, Addis AbabaUniversity, Addis Ababa, Ethiopia6Department of Clinical Pharmacology and Pharmacoepidemiology,University Hospital of Heidelberg, Heidelberg, Germany

Submitted for publication 1 March 2016; accepted 21 April 2016.

Corresponding Author:Abiy Habtewold,PhD,Department of Pharmaceutical Sciences, School ofPharmacy, 1050 Union University Dr, Jackson, TN 38305Email: abiyeyakem@gmail.com

2 The Journal of Clinical Pharmacology / Vol 00 No 0 2016

undergoes direct N-glucuronidation via UGT2B7.16,17

These metabolic pathways have been suggested toprevail in individuals with alleles of the reducedCYP2B6 variants, such as in the CYP2B6*6/*6genotype.14,17,18 Limited studies have attempted toexplore the roles of efflux and influx transporters intransmembrane dispositions of EFV. Those availablefor ABCB1 efflux transporters are conflicting.19,20

RIF has been demonstrated to induce the mainmetabolizing enzyme of EFV, CYP2B6. In primaryhuman hepatocytes, the increase in CYP2B6 activitybecause of RIF led to a 2.5-fold to 13-fold variationin EFV exposure.21,22 This was supported by in vivocoadministration of EFV with RIF that led to a mod-est (22%–26%) reduction in EFV plasma exposure inhealthy volunteers.1 Similarly, RIF was also reported toinduce the CYP3A and UGT2B7 enzymes.23 RIF is asubstrate and inducer of efflux drug transporters suchasABCB1,17,18 and it is also a substrate and inhibitor ofinflux transporters such as OATP1B1.24 These factorsmight substantially affect its pharmacokinetics (PK),which, in turn, may alter the PK of coadministereddrugs such as EFV.

Many of the previous studies that investigated theimpact of RIF-based anti-TB treatment analyzed only1 point concentration value of EFV, either Cmin orCmid-dose.11,25–27 These methods are more variable andless reliable than methods that determine PK parame-ters from multiple concentration values in a subject. Ofstudies that investigated the impact of RIF-based anti-TB treatment on PK parameters, a few involved healthyvolunteers1,28 or pediatric patients29 or did not considerpharmacogenetic information.30,31 A study by Bertrandet al used sparse PK samples from 10 subjects collected6 months after stopping RIF-based anti-TB treatment.These limited data are insufficient to capture the effectof coadministration.32

Two studies have reported the PK parameters of8-OH-EFV after a single dose of EFV in healthyvolunteers; only one of these studies investigatedthe effect of RIF on 8-OH-EFV.33,34 Neither studyreflected the underlying long-term situations in pa-tients who receive EFV-based combination antiretro-viral drugs and a RIF-based anti-TB regimen fora substantial duration. This is especially importantwhen there is renewed interest in investigating the PKparameters of the major metabolite of EFV, 8-OH-EFV partly because of its potential association withneurotoxicity.35

Therefore, the present study was designed to pri-marily investigate the effect of coadministration ofRIF-based anti-TB treatment on the PK parametersof EFV and its main metabolite, 8-hydroxy-efavirez (8-OH-EFV), using a model-independent approach.

MethodsExperimental DesignAll patients provided written informed consent toparticipate. The study protocol received approvalsfrom the Institutional Review Board (IRB) of theSchool of Medicine, Addis Ababa University, Na-tional EthicsReviewCommittee and subsequently fromthe Drug Administration and Control Authority ofEthiopia. The study also received approval from theIRB of Karolinska Institutet (Stockholm, Sweden)and was conducted per International Conference forHarmonization—Good Clinical Practice guidelines.This pharmacokinetic (PK) study was conducted asone of the substudies designed under the umbrellaof the broad clinical research named the HIV-TBPharmagene Study in Ethiopia, a sub-Saharan Africancountry. The patients were recruited from 4 study siteswithin Addis Ababa, the capital city of Ethiopia. Themain study site was Tikur Anbessa Specialized Refer-ral Hospital, and others were health centers. Detailsof the umbrella study design, the patient enrollmentprocess, and inclusion criteria with follow-up and drugtreatments were reported previously.10 Briefly, in thefull study, cohorts of newly diagnosed HIV-infected(n = 285) and TB-HIV-coinfected (n = 208) patientswith a baseline CD4 count less than 200 cells per cubicmillimeter were recruited in parallel and followed fora year. All patients in both groups received 600 mgof EFV daily as part of combination antiretroviraltherapy consisting of lamivudine (3TC) with eitherzidovudine (AZT) or stavudine (d4T) or tenofovir(TDF). In addition, the TB-HIV treatment group alsoreceived anti-TB treatment containing rifampicin +isoniazid + pyrazinamide + ethambutol, which wasinitiated 4 weeks prior to the start of EFV-basedcombination antiretroviral therapy.

Previous studies reported a 22%–26% reduction inEFV plasma exposure by RIF cotherapy.1,36,37 For thepresent PK substudy, a sample size of 20 subjects pertreatment group was considered optimal to detect a20%difference in the EFVplasma exposure between the2 treatment groups and 20% intraindividual variabilitywithin the TB-HIV treatment group (in the presenceand absence of RIF), with a sample power of 80% andα = 0.05 as the level of significance.38 As patients inthe TB-HIV treatment group were required to undergoPK sampling twice, more patients were enrolled inthis treatment group to account for loss to follow-up.From the HIV-only and TB-HIV treatment groups, 20and 33 patients, respectively, were randomly selectedto participate in the present PK substudy on the 16thweek of EFV-based combination antiretroviral therapyinitiation. In addition, the same patients from TB-HIVtreatment group were scheduled to undergo the second

Habtewold et al 3

Figure 1. A diagram showing the study design by treatment group.

PK sampling 8weeks after stoppingRIF-based anti-TBtreatment corresponding toweek 32. Figure 1 shows thetimeline of study design by treatment group.

Sample CollectionThe day before the PK sampling, each patient wasreminded to arrive at the study center fasted for atleast 6 hours and without swallowing their once-a-day evening EFV dose. The dose was administered inthe clinic a half hour prior to the start of sampling.In the evening of the sampling day, each patient wasadmitted for 24 hours at the study center. A cannulawas inserted into the brachial vein, and an 8-mL bloodsample was drawn into a CPT Vacutainer (BecktonDickinson, Heidelberg, Germany) 0, 1, 2, 3, 4, 6, 8,16, and 24 hours after the dose of EFV. The plasmasamples were immediately separated from blood cellswith ultracentrifugation for 10 minutes at 10 000 rpm.Plasma was aliquoted into cryotubes and stored at–80°C until shipment to Analytical Chemistry Labora-tory, Department of Clinical Pharmacology and Phar-macoepidemiology, University Hospital of Heidelberg(Germany) for bioanalysis.

Quantification of EFV and 8-OH-EFV ConcentrationsPlasma concentrations of EFV and 8-OH-EFV weredetermined by liquid chromatography coupled to tan-dem mass spectrometry, using a previously describedmethod.10,17,19,39,40 In brief, plasma proteins were pre-cipitated with ice-cold acetonitrile. The extract un-derwent chromatography on a Phenomenex SynergiFusion RP column with an eluent consisting of acid-ified 5 mM ammonium acetate buffer, acetonitrile,and methanol. EFV was quantified using 13C6-EFVas internal standard and negative electrospray tandem

mass spectrometry in the selected reaction monitor-ing mode. The lower limit of quantification (LLOQ)was 10 ng/mL. Linear regression with 1/X weightingresulted in correlation coefficients of R2 � 0.99. Theaccuracy and precision (intrabatch and interbatch) ofthe assay fulfilled all the recommendations of the USFood and Drug Administration bioanalytical methodvalidation guidelines. Quality control results for accu-racy and precision were within 20% of the nominalvalues including the LLOQ.

Estimation of PK ParametersPK parameters were determined by a model-independent method using WinNonlin professionalsoftware (version 6.4; Pharsight, Mountain View,California). The area under the plasma concentration–time curve (AUC0–24) was estimated using the lineartrapezoidal rule. The maximum and minimum plasmaconcentrations (Cmax, and Cmin) and the time tomaximum concentration (Tmax) were also determinedby visual inspection of the concentration-versus-timeplots for both EFV and 8-OH-EFV. Oral clearance(CL/F) for EFV was calculated as dose divided byAUC0–24.

CYP2B6 GenotypingGenomic DNA was isolated from peripheral bloodleukocytes using a QIAamp DNA Maxi Kit (QIA-GEN GmbH, Hilden, Germany). Genotyping for thecommon functional variant alleles in 5 relevant genesfor EFV disposition were carried out at the Divisionof Clinical Pharmacology, Department of LaboratoryMedicine, Karolinska Institutet, Stockholm, Sweden.Genotyping was done by real-time polymerase chainreaction (PCR) using predeveloped Taqman assay

4 The Journal of Clinical Pharmacology / Vol 00 No 0 2016

reagents for allelic discrimination (Applied BiosystemsGenotyping Assays). Allelic discrimination reactionswere performed using TaqManH (Applied Biosystems,Foster City, California) genotyping assays with theID number C__11711730_20 for the SNP CYP2B6*6c.516G.T ABI 7500 FAST (Applied Biosystems, FosterCity, California). The final volume for each reactionwas 10 mL, consisting of 2 × TaqMan UniversalPCR Master Mix (Applied Biosystems, Foster City,California), 20 × drug-metabolizing genotype assaymix, and 10 ng of genomic DNA. The PCR profileconsisted of an initial step at 50°C for 2 minutesand 50 cycles with 95°C for 10 minutes and 92°C for15 seconds.

Statistical AnalysesTo compare equivalence of PK exposure measureswith the 80%–125% criterion recommendation, 2 one-sided t tests were performed by constructing geomet-ric mean ratios and 90% confidence intervals (CIs)of “on” rifampicin as test treatment and “off” ri-fampicin as reference. The exposure parameters (exceptTmax) were log-transformed before statistical analyses.Multiple-group comparisons of exposure parametersfor EFV, 8-OH-EFV, and the metabolic ratio (EFV/8-OH-EFV) were done using 1-way analysis of vari-ance (ANOVA) followed by Bonferroni-corrected posthoc analysis. Independent t tests were used betweenindependent groups or subgroups to compare meansof the parameters, except the Tmax, for which thenonparametric Mann-Whitney U test was used. TheWilcoxon signed-rank test was used for the Tmax within-subject comparison.After analysis, the log-transformeddata were back-transformed and presented as geomet-ric means and 90%CIs. All statistical analyses wereperformed by Statistical Package for the Social Sci-ences (SPSS) software (version 16.0; SPSS, Chicago,Illinois), with α = 0.05. Plots were made using theggplot2 package in 3.2.3 version of R in the RStudioenvironment.

ResultsPatient CharacteristicsThe distribution of demographic and baseline clinicallaboratory characteristics is presented in Table 1, strat-ified by HIV and TB-HIV treatment groups, whereasFigures 2 and 3 show superimposed concentration-versus-time profiles of EFV and 8-OH-EFV, respec-tively, in the treatment groups. There were no statisticaldifference in demographic and baseline laboratorycharacteristics between the treatment groups. A sum-mary of geometric mean (90%CI) of PK exposuremeasures of EFV, 8-OH-EFV, and metabolic ratiosstratified by treatment group is presented in Table 2.

Table 1. Demographic Characteristics and Baseline Laboratory Valuesof the Patients

Treatment Group

Variables

HIVTreatment(n = 20)

TB-HIVCotreatment(n = 33)

I. Continuous variables Mean ± SD

Age (years) 34.5 ± 11.5 36.0 ± 8.5Body weight (kg) 53.0 ± 8.0 50.0 ± 7.2BMIa (kg/m2) 20.0 ± 2.6 18.6 ± 3.0Hemoglobin (g/dL) 12.5 ± 1.8 11.5.0 ± 2.2Leukocytes (×103 cells/μL) 4.6 ± 1.5 6.4 ± 2.8Aspartate aminotransferase (AST; U/L) 50.0 ± 41.8 50.5 ± 31.0Alanine aminotransferase (ALT, U/L) 38.8 ± 28.8 33.1 ± 19.0Alkaline phosphatase (ALP, U/L) 139.9 ± 79.6 146.8 ± 83.7Total bilirubin (mg/dL) 13.7 ± 6.8 11.9 ± 10.3Serum creatinine (mg/dL) 0.8 ± 0.2 0.9 ± 0.2HIV RNAb (log10 copies/mL) 5.2 ± 1.1 4.7 ± 1.1CD4

c (cells /mm3) 115.4 ± 58.9 96.5 ± 50.7

II. Categorical variables Counts in each subcategory

Sex (male, female) (14, 6) (13, 20)ARTd (zidovudine, stavudine, tenofovir) (9, 11, 0) (14, 14, 5)CYP2B6*6(516G>T) (GG, GT/TT) (7, 13) (13, 19)

aBMI, body mass index.bHIV RNA, ribonucleic acid of human immunodeficiency virus.cCD4, cluster of differentiation 4.dART, antiretroviral therapy plus (lamivudine + efavirenz).

Effect of RIF-Based Anti-TB Treatment on EFV and 8-OH-EFV PK ParametersOne-way ANOVA showed no difference in EFV ex-posure measures between the multiple groups, namely,only HIV infected and TB-HIV coinfected during the“on” and “off” RIF-based anti-TB regimen. However,significant differences between themultiple groups wereshown in AUC0–24 and Cmin of 8-OH-EFV and theirmetabolic ratios (Table 2). The Bonferroni-correctedpost hoc analysis revealed significant differences inAUC0–24 and Cmin of 8-OH-EFV and their metabolicratios between the HIV-only infected group and theTB-HIV-coinfected group, regardless of the presenceor absence of the RIF-based anti-TB regimen in thelatter group. The PK exposure measures of EFV and 8-OH-EFVwere compared between the treatment groups(HIV and TB-HIV; Figure 4). Parameters were alsocompared between the treatment groups with stratifi-cation based on sex (Table 3) and CYP2B6 genotypestatus (Table 4) in subgroup analyses.

No treatment group differences in EFV parameterestimates were noted without stratification (Figure 4).However, the TB-HIV-cotreated group showed a higherAUC 0–24 a and Cmin of 8-OH-EFV compared with theHIV-infected treatment group (Figure 4).

To control for the effect of sex, within-sex com-parisons of the parameters were performed between

Habtewold et al 5

Figure 2. Concentration-versus-time profiles of EFV showing no difference between the treatment groups. The red identity line with open circlesrepresent the mean profiles with individual observations for HIV treatment group on EFV, respectively,while the blue identity line with open rectanglesrepresent the mean profiles with individual observations for TB-HIV treatment group “on” EFV+RIF, respectively and the green identity line with opentriangles, represent the mean EFV profiles with individual observations for TB-HIV treatment group when “off” RIF, respectively.

treatment groups. No differences in EFV PK parame-ters were noted between the treatment groups in menand women (Table 2). In within-sex comparisons ofthe 8-OH-EFV parameters, women treated for TB-HIVhad greater AUC0–24 and Cmin compared with womentreated for only HIV; men treated for TB-HIV hadgreater Cmin compared with men treated for only HIV(Table 3).

Very few patients were CYP2B6*6/*6 homozygotes(3 patients in each treatment group); thus, patients weregrouped for statistical comparisons based on the pres-ence (carriers) or absence (noncarriers) of CYP2B6*6variant alleles. The effect of the RIF-based anti-TBregimen on EFV PK parameters between treatmentgroups is presented in Table 4. In the between-treatmentgroup comparisons, no difference in EFV PK param-eters was noted within the same CYP2B6 genotypecomparisons, except percent fluctuation was higher inCYP2B6*1/*1 genotypes than CYP2B6*6 carriers inthe TB-HIV treatment group, but not in the HIV-onlygroup (Table 4).However, increasedAUC0–24, Cmax, andCmin of 8-OH-EFV were observed when CYP2B6*6carriers in TB-HIV-cotreated patients were comparedwith CYP2B6*6 carriers in HIV-only treated patients(Table 4).

After stratifying by baseline bodyweight as �50 or>50 kg, EFV and 8-OH-EFV PK exposure measureswere compared between weight groups within each

treatment group. An independent t test showed nodifference in EFV and 8-OH-EFV exposure parametersbetween patients weighing �50 and >50 kg in bothtreatment groups.

Within-subject (intrasubject) differences in param-eters were evaluated when TB-HIV-cotreated patientswere “on” and “off” the RIF-based anti-TB regimen.The equivalence test of EFV AUC0–24, Cmax, andCmin were 100.5% (90%CI, 98.7%–102.3%), 100.2%(90%CI, 98.1%–102.4%), and 106.8% (90%CI, 97.6%–116.0%), respectively, whereas for 8-OH-EFV they were98.6% (90%CI, 95.5%–101.7%), 97.6% (90%CI, 92.2%–103.0%), and 101.1% (90%CI, 94.2%–107.9%), respec-tively. Neither EFV nor 8-OH-EFV showed differencesin concentrations and exposure parameters comparingbeing “on” and “off” the RIF-based anti-TB regimen(Figures 2, 3 and 4).

DiscussionThe present study showed no significant difference inthe parent (EFV) PK parameters between TB-HIV-cotreated patients and those treated for HIV only.A difference in metabolite (8-OH-EFV) exposure wassignificant, with TB-HIV cotreated patients havinghigher 8-OH-EFV exposure parameters compared withHIV-only patients. These findings may have practicalimportance in dosing decisions.

6 The Journal of Clinical Pharmacology / Vol 00 No 0 2016

Figure 3. Concentration-versus-time profiles of 8-OH-EFV in each treatment group. The red identity line with open circles represent the meanprofiles with individual observations for HIV treatment group on EFV, respectively, while the blue identity line with open rectangles represent themean profiles with individual observations for TB-HIV treatment group “on” EFV+RIF, respectively and the green identity line with open trianglesrepresent the mean 8-0H-EFV profiles with individual observations for TB-HIV treatment group when “off” RIF, respectively.

Table 2. Summary of PK Exposure Measures of EFV, 8-OH-EFV, and Metabolic Ratios Between the 3 Treatment Groups

Parameter, Geometric Mean(90%CI)a HIV Only, n = 20 TB-HIV “On” RIF, n = 33 TB-HIV “Off” RIF, n = 27c

AUC0–24 (ng · h/mL) 40 597 (34 095–48 339) 43 301 (34 809–53 851) 44 126 (36 199–53 777)Efavirenz Cmax (ng/mL) 3220 (2724–3806) 3265 (2713–3930) 3624 (3088–4253)

Cmin (ng/mL) 791 (595–1052) 1000 (748–1337) 821 (582–1160)CL/F (L/h) 14.8 (12.4–17.6) 13.9 (11.1–17.2) 13.6 (11.2–16.6)

8-Hydroxy-efavirenz AUC0–24 (ng · h/mL) 1386 (866–2217) 2568 (2182–3023) 2672 (1923–3713)Cmax (ng/mL) 121 (82–178) 170 (145–198) 202 (139–293)Cmin (ng/mL) 30 (22–41) 68 (57–81) 65 (52–81)

Metabolic ratio AUC0–-24 (ng · h/mL) 32.4 (24.6–42.6) 23.3 (21.9–24.9) 23.8 (20.7–27.4)(EFV/8-OH-EFV) Cmax (ng/mL) 55.9 (40.6–76.9) 39.7 (35.3–44.8) 39.7 (31.7–49.7)

Cmin (ng/mL) 99.8 (68.1–146.2) 47.6 (39.2–57.9) 44 (33.3–57.9)

Cmax, maximum concentrations; Cmin, minimum concentration; CL/F, apparent clearance; n, number of subjects.aExpressed as geometric mean (90% confidence interval).bSignificant with α < 0.05; AUC0–24, area under the curve of 0 –24 hours.cOf the 33 patients who participated in the first PK sampling, 6 declined to participate in the second PK sampling.

There was no significant difference in EFV expo-sure between the treatments in within-sex comparisons,although women receiving TB-HIV cotreatment hadhigher 8-OH-EFV parameters than women receivingonly HIV treatment. When controlling the effect forCYP2B6 genotype, again, there were no significantdifferences in EFV parameters between the treatmentswithin each CYP2B6 genotype, but CYP2B6*6 car-riers in the TB-HIV cotreatment group had signif-icantly higher 8-OH-EFV exposure parameters thantheir similar genotype counterparts in the HIV-only

treatment group. Within subjects, EFV and 8-OH-EFV exposure parameters were similar during andafter RIF-based anti-TB cotreatment with or withoutstratification based on sex or CYP2B6 genotype status.This was evident only in men, not in women.

In the present study, RIF-containing anti-TBcotreatment did not alter the PK parameters of EFVwithin TB-HIV-cotreated patients. This observationwas consistent in the within-subject comparisonsof EFV exposure parameters in TB-HIV-cotreatedpatients when they were “on” and “off” the RIF-based

Habtewold et al 7

Figure 4. Comparisons of PK parameters of EFV (left panel) and 8-OH-EFV (right panel) between treatment groups. EFV, Efavirenz; RIF, Rifampicin.

anti-TB treatment. In contrast to the parent drug,higher exposure of 8-OH-EFV (AUC0–24 increased79% and Cmin increased 105%) was noted in TB-HIV-cotreated patients compared with the HIV-onlytreated patients. This increased exposure to the 8-OH-EFV metabolite was not seen in the intrasubjectcomparisons with and without RIF of the cotreatedgroup. An expectation of increased 8-OH-EFVexposure parameters because pf RIF-based enzymeinduction might logically be accompanied by reductionin parent (EFV) exposure during RIF-based anti-TBtreatment. But this was not observed in the presentstudy. In fact, elevated exposure to 8-OH-EFV withno difference in EFV exposure was observed andpersisted for at least 8 weeks after the end of TB-HIVcotreatment. Reports on the washout period of RIFtreatment are not clear, ranging from 2 weeks to atleast 6 weeks. The present study adopted a conservativeapproach to allow 2 weeks more than the longestreported period, which was at least 6 weeks.28,41–43

Certainly, it would be challenging and is beyond thescope of this clinical study to prove the mechanisms

contributing to the above observation, given that EFVand 8-OH-EFV undergo both primary and secondarymetabolism via several enzymes, whose functionalitiesare affected by the presence of the anti-TB drugs andthe associated genetic polymorphisms in CYP2B6 andNAT2.44 However, the phenomena might be explainedthrough the following complementing mechanisms. Inpart, the presence of isoniazid (INH) in an anti-TB regimen might inhibit CYP2A6,44 redirecting themetabolism of EFV from 7-OH-EFV to 8-OH-EFVand contributing to the partial increase in the produc-tion of 8-OH-EFV. In addition, the continued long-term auto-induction and induction of CYP2B6 andCYP3A17,18,45,46 may lead to a further increase in 8-OH-EFV exposure.

However, when the anti-TB regimen was stopped,it would be expected that the inhibition of CYP2A6and induction of CYP2B6 and CYP3A invoked byINH and RIF, respectively, would have led to a de-crease in 8-OH-EFV exposure parameters, even withthe continued auto-inductions of CYP2B6 andCYP3Aby EFV.17,45,46 However, this was not observed in

8 The Journal of Clinical Pharmacology / Vol 00 No 0 2016

Table 3. Within-Sex, Between-Treatment Group Comparisons of EFV and 8-OH-EFV PK Parameters

Efavirenz, Geometric Mean (90%CI)a8-Hydroxy-Efavirenz, Geometric Mean

(90%CI)a

Parameters Sex HIV-Only Group TB-HIV Group P HIV-Only Group TB-HIV Group P

AUC0–24 (ng · h/mL) Male 47 973 (36 864–62 431) 49 728 (35 027–70 583) .84 2161 (1276–3659) 3141(2370–4163) .13n 14 13 14 13

Female 29 868 (25 038–35 629) 41 764 (30 960–56 351) .26 1117 (620–2012) 2382 (1959–2897) .01c

n 6 20 6 20

Cmax (ng/mL) Male 3654 (2815–4745) 3672 (2642–5103) .58 173 (101–294) 218 (159–299) .25n 14 13 14 13

Female 2476 (2031–3017) 3249 (2605–4052) .21 87 (52–144) 148 (126–175) .03c

n 6 20 6 20

Tmax (h) Male 3.7 (2.6–5.1) 4.4 (3.0–6.6) .43 4.2 (2.4–7.3) 3 (1.8–4.9) .35n 13 13 13 13

Female 3.8 (1.7–8.6) 3.6 (2.6–4.9) .86 2.9 (1.8–4.7) 3.1 (2.0–4.9) .83n 6 19 6 19

Cmin (ng/mL) Male 1128 (887–1434) 1262 (840–1897) .95 35 (22–56) 76 (58–99) .01c

n 14 13 14 13Female 451 (228–894) 980 (673–1425) .06 26 (15–46) 62 (48–80) .01c

n 6 20 6 20

CL/F (L/h) Male 12.5 (9.6–16.2) 12.0 (8.5–17.1) .91 NAb NA NAn 13 13

Female 20.0 (16.8–23.9) 14.3 (10.6–19.3) .29 NA NA NAn 6 19

AUC0–24, area under the curve of 0–24 hours; Cmax, maximum concentrations; Tmax, time to maximum concentration; Cmin, minimum concentration; CL/F,apparent clearance; n, number of subjects.aExpressed as geometric mean (90% confidence interval).bNA, not available.cSignificant with p < 0.05.

the present study, as there were no differences in 8-OH-EFV exposure parameters 8 weeks after stoppingthe RIF-based anti-TB regimen coadministration withEFV in TB-HIV-cotreated patients. Moreover, the un-derlying auto-induction by EFV and induction by RIFof CYP2B6 could have also resulted in reduced 8-OH-EFV exposure, as this enzyme catalyzes not only themetabolism of EFV to 8-OH-EFV but also 8-OH-EFVto 8,14-di-OH-EFV.14

Therefore, multiple mechanisms might seem to beinvolved in explaining the increase in 8-OH-EFV PKexposure measures in the presence of RIF without anychange in the parent (EFV) exposure measures. Oneplausible mechanism is mechanism-based inhibition ofenzymes by a reactive metabolite. The primary EFVmetabolite, 8-OH-EFV, was demonstrated to show suchinhibition on CYP2B6 in vitro.47 Again, one mightalso assume this is a vicious cycle, in that inhibition ofCYP2B6 by 8-OH-EFVmay also be expected to inhibitthe conversion of EFV to 8-OH-EFV, which maylead to enhanced exposure to EFV with a subsequentdecrease in 8-OH-EFV exposure because of reducedproduction. One reason this was not observed may bethat metabolism of EFV to 8-OH-EFV can still occur

via other enzymatic pathways, notably by CYP3A andCYP1A2.14

The present finding of the absence of the effectof RIF-based anti-TB cotreatment on EFV exposureparameters corroborates a number of previous reports.In a relatively similar study design in south Indianpatient populations, Ramachandran et al reported thateven though Cmax and AUC0–24 seemed to slightly butinsignificantly decrease with RIF-based cotreatment,they concluded that RIF coadministration did notsignificantly alter the PK parameters of EFV.48 Semvuaet al reported a similar outcome in Tanzanian patients,although Cmax, Cmin, and AUC0–24 were slightly butinsignificantly lower during RIF cotreatment.31 Noeffects of RIF coadministration on plasma concen-trations of EFV at Cmid-dose or Cmin were reportedin Cambodian, South African, Thai, Ugandan, andTanzanian patients.3,11,27,40,49,50 However, our finding isin disagreement with a recent report by Cho et al, whodemonstrated a decrease in EFV exposure parameterswith a corresponding increase in 8-OH-EFV on a singledose of EFV after 10 days of RIF pretreatment inhealthy volunteers,34 and with an earlier report byLopez-Cortes et al, who showed significant decreases

Habtewold et al 9

Table 4. Within-CYP2B6 Genotype, Between-Treatment Group Comparisons of EFV and 8-OH-EFV PK Parameters

Efavirenz, Geometric Mean (90%CI)a8-Hydroxy-Efavirenz, Geometric Mean

(90%CI)a

Parameters CYP2B6 Genotype HIV-Only Group TB-HIV Group P HIV-Only Group TB-HIV Group P

AUC0–24 (ng · h/mL) *1/*1 38 495 (29 861–49 625) 34 987(27 485–44 535) .55 3021 (1176–7766) 2739 (2017–3718) .88n 7 13 7 13

*6 Carrier 41 773 (31 089–56 131) 53 555 (37 558–76 366) .51 1255 (890–1771) 2489 (2068–2995) .02c

n 13 19 13 19

Cmax (ng/mL) *1/*1 3046 (2187–4241) 2971 (2449–3606) .79 258 (101–661) 187 (138–252) .64n 7 13 7 13

*6 Carrier 3264 (2507–4249) 3816 (2841–5126) .77 95 (71–128) 159 (131–193) .01c

n 13 19 13 19

Tmax (h) *1/*1 3.6 (3.1–4.3) 3.7 (2.5–5.4) .98 5.1 (3.0–8.9) 2.5 (1.5–4.2) .08n 6 13 6 13

*6 Carrier 3.7 (2.4–5.8) 4.01 (2.9–5.6) .78 3.19 (1.87–5.5) 3.5 (2.3–5.4) .79n 13 18 13 18

Cmin (ng/mL) *1/*1 898 (646–1248) 797 (591–1077) .61 51 (25–106) 70 (51–96) .35n 7 13 7 13

*6 Carrier 775 (479–1253) 1358 (881–2094) .25 24 (18–34) 62 (49–78) .001c

n 13 19 13 19

CL/F (L/h) *1/*1 15.5 (12.0–20.0) 17.1 (13.4–21.8) .55 NAb NA NAn 7 13

*6 Carrier 14.3 (10.6–19.2) 11.2 (7.8–15.9) .45 NA NA NAn 12 18

AUC(0-24), area under the curve of 0 – 24h;Cmax,maximum concentrations; Tmax, time to maximum concentration; Cmin,minimum concentration; CL/F, apparentclearance; N, number of subjects.aExpressed as geometric mean (90% confidence interval).bNA, not available.cSignificant with p < 0.05.

in Cmax, Cmin, and AUC0–24 in the presence of RIFcoadministration.1

The present study also compared EFV and 8-OH-EFV PK parameters between the HIV-only and TB-HIV-cotreatment groups while controlling for the effectof sex. In the present work, within-sex comparisonsshowed no difference between groups on EFV exposurein either sex. However, higher 8-OH-EFV exposurewas observed in TB-HIV-cotreated women, suggestingthe presence of a sex effect in the exposure of 8-OH-EFV, which was enhanced by the presence of RIF-based anti-TB therapy. Reports implying the effect ofsex in determining EFV PK are conflicting, such thatsome investigators showed higher exposure parametersin females than in males, whereas others demonstratedlower exposure, and a few reported no difference, asreviewed by Ofotokun et al and Roger et al.51,52 Itis not clear why these discrepancies were reported.To our knowledge, this is the first report describingthe effect of sex on 8-OH-EFV PK parameters. Infurther exploratory analysis, we screened whether thisdifference could be attributed to any of the baselinecharacteristics; nevertheless, none could be ascribed.On the other hand, there was no within-sex differencein EFV and 8-OH-EFV exposures when the TB-HIV-

cotreated subjects were “on”and “off”RIF-based anti-TB treatment.

The effect of CYP2B6 genotype was controlled tocompare EFV and 8-OH-EFV exposure parametersbetween the HIV-only and TB-HIV treatment groups.Even though differences in EFV exposure parame-ter were not observed between the treatment groupsconsidering both CYP2B6*1/*1 (fully functional) and*6 carriers (less functional) groups separately, 8-OH-EFV exposure parameters were higher in TB-HIVcotreatment than in the HIV-only group in *6 car-riers but not in CYP2B6*1/*1. As articulated above,EFV is metabolized by a number of enzymes suchas CYP2B6, CYP3A, CYP2A6, CYP1A2, and UGTenzyme systems, whereas 8-OH-EFV has CYP2B6 andUGT enzymes for its clearance. When CYP2B6 is lessfunctional, as in the case of *6 carriers, EFV wouldhave other clearance pathways to 8-OH-EFV,14,47 but8-OH-EFV would have only UGT, which might leave8-OH-EFV, but not EFV, exposure parameters higherin CYP2B6*6 carriers. In addition, 8-OH-EFV wasshown in vitro to cause mechanism-based irreversibleinhibition of CYP2B6, which would further reducethe already less functional enzyme in the CYP2B6*6carriers.47

10 The Journal of Clinical Pharmacology / Vol 00 No 0 2016

The equivalence of PK exposure measures waswithin the recommended 80%–125% confidence inter-val criterion53 for both EFV and 8-OH-EFV in within-subject comparisons, when the TB-HIV-cotreatedsubjects were “on” and “off” a RIF-based anti-TBregimen. This demonstrated equivalence suggests thata RIF-based anti-TB regimen does not influence PKmeasures of both EFV and 8-OH-EFV. Although thewithin-subject exposure measure comparisons of EFVand 8-OH-EFV showed no difference, a difference inthe exposure measures of 8-OH-EFV but not EFVwas exhibited in comparing HIV-only infected andTB-HIV-coinfected groups. This may be bercause ofthe group differences influencing the pharmacokineticsof 8-OH-EFV but not EFV. In an exploratory analy-sis, TB-HIV-coinfected groups had significantly higherbaseline leukocyte counts than the HIV-only infectedgroup. It is not clear how the baseline difference inleukocyte counts or the presence of TB coinfection withHIV or other unidentified characteristics between thegroups would be attributed to the differences in 8-OH-EFV PK while not affecting EFV concentrations.

Weight-based dosing of EFV in patients cotreatedfor TB-HIV has recently been advocated, stratified byeither less or greater than 50 or 60 kg, depending on theavailable guidelines.54,55 Although it is hard to concludegiven the sample size of the present study, this studydemonstrated the absence of a body weight effect onEFV and 8-OH-EFV PK exposure measures, as therewas no difference between the TB-HIV cotreatment andHIV-only treatment groups.

A limitation of the present study was the presence ofa very few patients with the homozygousCYP2B6 *6/*6allele. Although this was not the primary objective ofthe study, it might limit the interpretability of the resultsassociated with CYP2B6*6 genotypes.

ConclusionsLong-term coadministration of a RIF-based anti-TBregimen did not alter EFV PK parameters betweensubjects or within subjects, even after controlling forthe effect of either sex or CYP2B6 genotypes. Con-versely, coadministration did significantly increase thePK exposure parameters of the primary metabolite, 8-OH-EFV, an effect that was pronounced in women andCYP2B6*6 carriers. Hence, consistent with our recentreport on sparse data,10 the present study does notsuggest the need for an EFV dose increase from 600to 800 mg per day when EFV and RIF-based therapiesare coadministered. Given the pharmacogenetically di-verse nature of the sub-Saharan African population,39

further studies are warranted in other races exploringthe long-term effect of RIF coadministration on EFVand 8-OH-EFV exposure measures, neurotoxicities and

other untoward effects of higher exposures to 8-OH-EFV during simultaneous TB-HIV cotreatment, espe-cially in females and CYP2B6*6 carriers.

AcknowledgmentsWe are grateful to Lilleba Bohman for dedicated technicalassistance.

FundingThis work was supported by research grants from Europeanand developing countries clinical trial partnership (grantnumber CT.2005.32030.001); Swedish International develop-ment agency (grant numbers HIV-2006-031, SWE 2007–270),and Swedish research council (grant number 348-2011-7383).

Declaration of Conflicting InterestsNone to declare.

References1. Lopez-Cortes LF, Ruiz-Valderas R, Viciana P, et al. Phar-

macokinetic interactions between efavirenz and rifampicin inHIV-infected patients with tuberculosis. Clin Pharmacokinet.2002;41(9):681–690.

2. Matteelli A, Regazzi M, Villani P, et al. Multiple-dosepharmacokinetics of efavirenz with and without the use ofrifampicin in HIV-positive patients. Curr HIV Res. 2007;5(3):349–353.

3. Borand L, Laureillard D,Madec Y, et al. Plasma concentrationsof efavirenz with a 600 mg standard dose in Cambodian HIV-infected adults treated for tuberculosis with a body weight above50 kg. Antivir Ther. 2013;18(3):419–423.

4. Cohen K, Grant A, Dandara C, et al. Effect of rifampicin-basedantitubercular therapy and the cytochrome P450 2B6 516G>Tpolymorphism on efavirenz concentrations in adults in SouthAfrica. Antivir Ther. 2009;14(5):687–695.

5. Friedland G, Khoo S, Jack C, Lalloo U. Administration ofefavirenz (600 mg/day) with rifampicin results in highly variablelevels but excellent clinical outcomes in patients treated for tu-berculosis and HIV. J Antimicrob Chemother. 2006;58(6):1299–1302.

6. Gengiah TN, Botha JH, Yende-Zuma N, Naidoo K, KarimSS. Efavirenz dosing: influence of drug metabolizing enzymepolymorphisms and concurrent tuberculosis treatment. AntivirTher. 2015;20(3):297–306.

7. Kwara A, Ramachandran G, Swaminathan S. Dose adjustmentof the non-nucleoside reverse transcriptase inhibitors duringconcurrent rifampicin-containing tuberculosis therapy: one sizedoes not fit all. Expert Opin Drug Metab Toxicol. 2010;6(1):55–68.

8. Lee KY, Lin SW, Sun HY, et al. Therapeutic drug monitoringand pharmacogenetic study of HIV-infected ethnic Chinesereceiving efavirenz-containing antiretroviral therapy with orwithout rifampicin-based anti-tuberculous therapy. PLoS One.2014;9(2):e88497.

9. Lopez-Cortes LF, Ruiz-Valderas R, Ruiz-Morales J, et al.Efavirenz trough levels are not associated with virological fail-ure throughout therapy with 800 mg daily and a rifampicin-containing antituberculosis regimen. J Antimicrob Chemother.2006;58(5):1017–1023.

Habtewold et al 11

10. Habtewold A, Makonnen E, Amogne W, et al. Is there a needto increase the dose of efavirenz during concomitant rifampicin-based antituberculosis therapy in sub-Saharan Africa? The HIV-TB pharmagene study. Pharmacogenomics. 2015:1–17.

11. Orrell C, Cohen K, Conradie F, et al. Efavirenz and rifampicinin the South African context: is there a need to dose-increaseefavirenz with concurrent rifampicin therapy? Antiviral Ther.2011;16(4):527–534.

12. Sathia L, Obiorah I, Taylor G, et al. Concomitant use of nonnu-cleoside analogue reverse transcriptase inhibitors and rifampicinin TB/HIV type 1-coinfected patients. AIDS Res Hum Retrovir.2008;24(7):897–901.

13. Ward BA, Gorski JC, Jones DR, Hall SD, Flockhart DA, DestaZ. The cytochrome P450 2B6 (CYP2B6) is the main catalystof efavirenz primary and secondary metabolism: implicationfor HIV/AIDS therapy and utility of efavirenz as a substratemarker of CYP2B6 catalytic activity. J Pharmacol Exp Ther.2003;306(1):287–300.

14. Avery LB, VanAusdall JL, Hendrix CW, Bumpus NN. Com-partmentalization and antiviral effect of efavirenz metabolitesin blood plasma, seminal plasma, and cerebrospinal fluid. DrugMetab Dispos. 2013;41(2):422–429.

15. Ogburn ET, Jones DR, Masters AR, Xu C, Guo Y, Desta Z.Efavirenz primary and secondary metabolism in vitro and invivo: identification of novel metabolic pathways and cytochromeP450 2A6 as the principal catalyst of efavirenz 7-hydroxylation.Drug Metab Dispos . 2010;38(7):1218–1229.

16. Cho DY, Ogburn ET, Jones D, Desta Z. Contribution of N-glucuronidation to efavirenz elimination in vivo in the basal andrifampin-induced metabolism of efavirenz. Antimicrob AgentsChemother. 2011;55(4):1504–1509.

17. Habtewold A, Amogne W, Makonnen E, et al. Long-termeffect of efavirenz autoinduction on plasma/peripheral bloodmononuclear cell drug exposure and CD4 count is influencedby UGT2B7 and CYP2B6 genotypes among HIV patients.J Antimicrob Chemother. 2011;66(10):2350–2361.

18. Baciewicz AM, Chrisman CR, Finch CK, Self TH. Updateon rifampin and rifabutin drug interactions. Am J Med Sci.2008;335(2):126–136.

19. Burhenne J, Matthee AK, Pasakova I, et al. No evidence forinduction of ABC transporters in peripheral bloodmononuclearcells in humans after 14 days of efavirenz treatment. AntimicrobAgents Chemother. 2010;54(10):4185–4191.

20. Mukonzo JK, Roshammar D, Waako P, et al. A novel polymor-phism in ABCB1 gene, CYP2B6*6 and sex predict single-doseefavirenz population pharmacokinetics in Ugandans. Br J ClinPharmacol. 2009;68(5):690–699.

21. Faucette SR, Wang H, Hamilton GA, et al. Regulation ofCYP2B6 in primary human hepatocytes by prototypical induc-ers. Drug Metab Dispos. 2004;32(3):348–358.

22. Madan A, Graham RA, Carroll KM, et al. Effects of pro-totypical microsomal enzyme inducers on cytochrome P450expression in cultured human hepatocytes. Drug Metab Dispos.2003;31(4):421–431.

23. Gallicano KD, Sahai J, Shukla VK, et al. Induction of zidovu-dine glucuronidation and amination pathways by rifampicin inHIV-infected patients. Br J Clin Pharmacol. 1999;48(2):168–179.

24. Kalliokoski A, Niemi M. Impact of OATP transporters onpharmacokinetics. Br J Pharmacol. 2009;158(3):693–705.

25. Borand L,MadecY, LaureillardD, et al. Plasma concentrations,efficacy and safety of efavirenz in HIV-infected adults treatedfor tuberculosis in Cambodia (ANRS 1295-CIPRA KH001CAMELIA trial). PLoS One. 2014;9(3):e90350.

26. Kwara A, Lartey M, Sagoe KW, Court MH. Paradoxicallyelevated efavirenz concentrations inHIV/tuberculosis-coinfected

patients with CYP2B6 516TT genotype on rifampin-containingantituberculous therapy. AIDS. 2011;25(3):388–390.

27. Uttayamakul S, Likanonsakul S, Manosuthi W, et al. Effects ofCYP2B6 G516T polymorphisms on plasma efavirenz and nevi-rapine levels when co-administered with rifampicin in HIV/TBco-infected Thai adults. AIDS Res Ther. 2010;7:8.

28. Kwara A, Cao L, Yang H, et al. Factors associated with vari-ability in rifampin plasma pharmacokinetics and the relationshipbetween rifampin concentrations and induction of efavirenzclearance. Pharmacotherapy. 2014;34(3):265–271.

29. McIlleron HM, Schomaker M, Ren Y, et al. Effects ofrifampin-based antituberculosis therapy on plasma efavirenzconcentrations in children vary by CYP2B6 genotype. AIDS.2013;27(12):1933–1940.

30. Kwara A, Tashima KT, Dumond JB, et al. Modest but variableeffect of rifampin on steady-state plasma pharmacokinetics ofefavirenz in healthy African-American and Caucasian volun-teers. Antimicrob Agents Chemother. 2011;55(7):3527–3533.

31. Semvua HH, Mtabho CM, Fillekes Q, et al. Efavirenz, tenofovirand emtricitabine combinedwith first-line tuberculosis treatmentin tuberculosis-HIV-coinfected Tanzanian patients: a pharma-cokinetic and safety study. Antivir Ther. 2013;18(1):105–113.

32. Bertrand J, Verstuyft C, ChouM, et al. Dependence of efavirenz-and rifampicin-isoniazid-based antituberculosis treatment drug-drug interaction on CYP2B6 andNAT2 genetic polymorphisms:ANRS 12154 study in Cambodia. J Infect Dis. 2014;209(3):399–408.

33. Abdelhady AM, Desta Z, Jiang F, Yeo CW, Shin JG, Over-holser BR. Population pharmacogenetic-based pharmacokineticmodeling of efavirenz, 7-hydroxy- and 8-hydroxyefavirenz. J ClinPharmacol. 2014;54(1):87–96.

34. Cho DY, Shen JHQ, Lemler SM, et al. Rifampin en-hances cytochrome P450 (CYP) 2B6-mediated efavirenz 8-hydroxylation in healthy volunteers.DrugMetab Pharmacokinet.2016;31(2):107–116.

35. Tovar-y-Romo LB, Bumpus NN, Pomerantz D, et al. Dendriticspine injury induced by the 8-hydroxy metabolite of efavirenz.J Pharmacol Exp Ther. 2012;343(3):696–703.

36. Efavirenz package insert. http://packageinsertsbmscom/pi/pi_sustivapdf. Accessed April 2016.

37. Csajka C, Marzolini C, Fattinger K, et al. Population phar-macokinetics and effects of efavirenz in patients with hu-man immunodeficiency virus infection. Clin Pharmacol Ther.2003;73(1):20–30.

38. Kang D, Schwartz JB, Verotta D. Sample size computationsfor PK/PD population models. J Pharmacokinet Pharmacodyn.2005;32(5-6):685–701.

39. Ngaimisi E, Habtewold A, Minzi O, et al. Importance ofethnicity, CYP2B6 and ABCB1 genotype for efavirenz pharma-cokinetics and treatment outcomes: a parallel-group prospectivecohort study in two sub-Saharan Africa populations. PLoS One.2013;8(7):e67946.

40. Ngaimisi E, Mugusi S, Minzi O, et al. Effect of rifampicin andCYP2B6 genotype on long-term efavirenz autoinduction andplasma exposure in HIV patients with or without tuberculosis.Clin Pharmacol Ther. 2011;90(3):406–413.

41. Humbert G, Brumpt I, Montay G, et al. Influence of rifampinon the pharmacokinetics of pefloxacin. Clin Pharmacol Ther.1991;50(6):682–687.

42. Martin P, Oliver S, Robertson J, Kennedy SJ, Read J, Du-vauchelle T. Pharmacokinetic drug interactions with vandetanibduring coadministration with rifampicin or itraconazole. DrugsR D. 2011;11(1):37–51.

43. Park JY, Kim KA, Park PW, Park CW, Shin JG. Effect ofrifampin on the pharmacokinetics and pharmacodynamics ofgliclazide. Clin Pharmacol Ther. 2003;74(4):334–340.

12 The Journal of Clinical Pharmacology / Vol 00 No 0 2016

44. Luetkemeyer AF, Rosenkranz SL, Lu D, et al. Combined effectof CYP2B6 and NAT2 genotype on plasma efavirenz exposureduring rifampin-based antituberculosis therapy in the STRIDEstudy. Clin Infect Dis. 2015;60(12):1860–1863.

45. Fletcher CV, Brundage RC, Fenton T, et al. Pharmacokineticsand pharmacodynamics of efavirenz and nelfinavir in HIV-infected children participating in an area-under-the-curve con-trolled trial. Clin Pharmacol Ther. 2008;83(2):300–306.

46. McDonagh EM, Lau JL, Alvarellos ML, Altman RB, Klein TE.PharmGKB summary: efavirenz pathway, pharmacokinetics.Pharmacogenet Genomics. 2015;25(7):363–376.

47. Bumpus NN, Kent UM, Hollenberg PF. Metabolism ofefavirenz and 8-hydroxyefavirenz by P450 2B6 leads to inac-tivation by two distinct mechanisms. J Pharmacol Exp Ther.2006;318(1):345–351.

48. Ramachandran G, Hemanth Kumar AK, Rajasekaran S,et al. CYP2B6 G516T polymorphism but not rifampincoadministration influences steady-state pharmacokineticsof efavirenz in human immunodeficiency virus-infected patientsin South India. Antimicrob Agents Chemother. 2009;53(3):863–868.

49. Mukonzo JK, Bisaso RK, Ogwal-Okeng J, Gustafsson LL,Owen JS, Aklillu E. CYP2B6 genotype-based efavirenz dose rec-ommendations during rifampicin-based antituberculosis cotreat-

ment for a sub-Saharan Africa population. Pharmacogenomics.2016;17(6):603–613.

50. Mukonzo JK, Nanzigu S, Waako P, Ogwal-Okeng J, GustafsonLL, Aklillu E. CYP2B6 genotype, but not rifampicin-basedanti-TB cotreatments, explains variability in long-term efavirenzplasma exposure. Pharmacogenomics. 2014;15(11):1423–1435.

51. Ofotokun I, Chuck SK,Hitti JE. Antiretroviral pharmacokineticprofile: a review of sex differences.GendMed. 2007;4(2):106–119.

52. Rotger M, Csajka C, Telenti A. Genetic, ethnic, and genderdifferences in the pharmacokinetics of antiretroviral agents.CurrHIV/AIDS Rep. 2006;3(3):118–125.

53. U.S. Food and Drug Administration C. Drug InteractionStudies — Study Design, Data Analysis, Implications forDosing, and Labeling Recommendations (2012). http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm292362pdf. Accessed February 21,2016.

54. U.S. Food and Drug Administration. Sustiva labelingupdate / dosing adjustment with rifampin. http://archiveis/Qqaq6#selection-5110-51157. February 2012. Accessed April2016.

55. Pozniak AL, Coyne KM, Miller RF, et al. British HIV Associ-ation guidelines for the treatment of TB/HIV coinfection 2011.HIV Med. 2011;12(9):517–524.