Open label pharmacokinetic RBT dose-finding study in young (≤ 5 year old) HIV-infected children receiving a LPV/RTV-based ART regimen and who have a recent history of completing TB treatment.

BACKGROUND AND RATIONALE The HIV and TB epidemics disproportionately affect sub-Saharan Africa, and have had a catastrophic impact in the region. The synergy between these pathogens has resulted in dramatic increases in TB incidence, with 58% of cases of adult TB in 2005 in South Africa being HIV-associated [1]. The impact of HIV co-infection on TB in children has not yet been well defined due to inadequate surveillance and diagnostic challenges of pediatric TB and HIV in resource poor settings. Studies have demonstrated that between 10 and 60% of children with TB in sub-Saharan Africa are co-infected with HIV [2-5]. The risk of diagnosing TB among HIV-infected children has been shown to be 8 times higher among HIV-infected children than among HIV-uninfected children [6]. The risk of TB among HIV-infected children is associated with the degree of immunodeficiency, with a 3-year risk of developing TB among immunocompromised children (CD4<15%) of 12,400/100,000, compared to 3,300/100,000 in less immunocompromised HIV-infected children [7]. TB in HIV infected children is not only a cause of morbidity, but is also a frequent cause of death among HIV-infected children [8].

Even though the treatment of HIV-associated TB is complicated by overlapping toxicities, high pill burden and the occurrence of immune reconstitution inflammatory syndrome (IRIS), experiences of concomitant TB and HIV treatment in children are positive, with HIV-infected children with active TB who start ART having similar outcomes to children who started on ART for a non-TB diagnosis [9].

Co-administration of TB and HIV medications is associated with clinically significant and highly complex drug-drug interactions. Protease inhibitors (PIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs) are substrates for cytochrome P450 isozymes and P-glycoprotein. Rifampin (RIF), a key component of TB treatment regimens, is a potent inducer of hepatic and intestinal cytochrome P450 isozymes and P-glycoprotein [10]. Multiple studies have demonstrated decreased concentrations of PIs and NNRTIs when coadministered with rifampin [11]. The concentrations of Efavirenz (EFV), nevirapine (NVP) and LPV/RTV are decreased by as much as 22, 68 and 75%, respectively, when coadministered with RIF [12, 13]. While the interaction with EFV may not be clinically important or could be overcome by dose increase [12], investigations with LPV/RTV have not been able to provide full confidence that higher doses of LPV/RTV overcome induction of enzymes by RIF [14].

In resource-poor countries, most adults with HIV/TB co-infection receive a combination of EFV-based ART and RIF-based TB treatment. In children under 3 years of age, EFV is not recommended for antiretroviral treatment. The choice of ART regimen in TB/HIV co-infected children is further complicated by the limited options for pediatric drug formulations and/or dosing information. According to the World Health Organization guidelines, the first-line treatment recommendation for children with TB and HIV co-infection is a triple NRTI regimen consisting of d4T or AZT plus 3TC and ABC [15]. Alternatively, the standard first-line regimen of two NRTIs plus NVP can be used. Both of these regimens are suboptimal. The triple NRTI regimen has been shown to have suboptimal virologic potency [16], and the NVP-based regimen is problematic in children exposed to single dose NVP for prevention of PMTCT as many of these will have developed resistance [17].

In South Africa, the recommended regimen for children age < 3 years of age who receive concomitant treatment for TB is the combination of LPV/RTV with an NRTIs. A recent study demonstrated that this regimen also results in suboptimal doses of LPV/TRV when given concomitantly with RIF, and that even a double dose of LPV/RTV given concomitantly with RIF does not provide adequate LPV exposures [18] . Another study assessed super-boosted LPV as an alternative approach, using LPV/RTV with extra RTV to equal the dose of LPV but this study was stopped due to safety concerns [19].

Clearly, there is an urgent need for a safe, effective and feasible combined treatment for young children with HIV-associated TB. Rifabutin (RBT), a wide spectrum antibiotic particularly active on acid-fast bacilli is a rifamycin with limited CYP450 induction potential and thus minimal effect on concentrations of NNRTIs and PIs including LPV/RTV [20], is a recommended alternative to treat TB in combination with PI-based ART in adults [21]. In adults, RBT has been found to be equally safe and effective as RIF for the treatment of TB, with similar rates of adverse events, cure and relapse of TB between RIF and RBT [22]. The information from randomized clinical trials is however dominated by HIV negative individuals The only comparative randomized clinical trials in HIV positive patients found both RIF and RBT to be safe and effective and demonstrated more rapid clearance of acid-fast bacilli when using RBT [23]. The evidence from observational cohort studies including HIV-infected patients treated with ART does not point to inferior performance of rifabutin, but is less rigorous. To date, there are no published efficacy studies of RBT in combination with these boosted PIs. In addition, the clinical experience with RBT for TB in resource-constrained settings is limited, especially in the context of the PIs [24].

In March 2009, Rifabutin (RBT) was added to the WHO essential medicine list to substitute RIF for treatment of TB in HIV infected patients treated with RTV boosted PI-containing antiretroviral therapy [25]. Listing RBT on the essential medicine list, was an important first step toward increasing the availability of RBT for large scale use at affordable cost in the resource-limited settings. The high cost of RBT when compared with RIF however made it inaccessible to tuberculosis control programs in resource-limited settings. In August 2009, the Clinton Foundation announced a price agreement with Pfizer to lower the price of RBT in emerging markets by 60%, to $1 per 150mg dose. This initiative greatly improved the prospect of using RBT in a public health approach to co-treatment of TB and HIV in developing countries.

The major remaining stumbling block to use RBT for concomitant treatment of TB and HIV in children is the lack of data (dosing, safety and efficacy) on the use of RBT for treatment of TB in children. Until now, dosing for RBT is only available to treat Mycobacterium Avium Complex (MAC) in children. Proper dosing of RBT in combination with LPV/TRV ART is essential to prevent failure of therapy and rise of drug resistant TB and HIV strains.

The proposed assessment of the optimal dosing strategy, pharmacokinetic profile, and safety of rifabutin when given concomitantly with LPV/RTV in young children (age < 5 years) will be a critical step towards improved management of TB/HIV in these children. Availability of a safe and effective treatment regimen for young HIV co-infected children with TB is highly important, especially in light of the results of the Children with HIV Early Antiretroviral Therapy (CHER) trial which indicate that HIV-infected infants should be treated with ART as early as possible (in the first months of life) as this dramatically decreases their risk of early death [26].

HYPOTHESIS AND STUDY OBJECTIVES

Hypothesis We hypothesize that a clinically significant drug interaction exists between rifabutin (RBT) and lopinavir/ritonavir (LPV/RTV) in young children (age < 5 years) such that rifabutin dosing for concomitant administration of RBT and LPV/RTV needs to be developed

Primary Objective To assess dosing, pharmacokinetic profile, and safety of rifabutin when given concomitantly with LPV/RTV for 14 days in HIV-infected children age < 5 years.

STUDY PROCEDURES Rifabutin administration, formulation and preparation Rifabutin (RBT) will be administered at a dose of 5mg/kg three times per week for a period of 14 days in 10 children. All six RBT dosages will be taken under direct observation, witnessed at the subject's home (community based DOT). The amount of drug dispensed and returned will also be recorded by the pharmacy. A diary card will be provided to all parents and guardians to record dosage administration (adherence) of all other drugs throughout the study.

RBT will be compounded at the Baragwanath hospital using 150mg Mycobutin capsules. Rifabutin suspension will be provided by the study as a 14 day supply. While there is currently no commercially available RBT dosage form for use in children, rifabutin has been compounded as a suspension for easy use in children and was found stable for 12 weeks at 4 to 40 °C [27].

Since we anticipate minimal change in LPV exposures with the addition of RBT (AUC ratio of LPV+RBT/ LPV alone is 1.17, 90%CI 1.04-1.31) [28], LPV/RTV dose will not be altered during co-administration with RBT.

Initial Dose Justification There is no PK data in children for using RBT alone or in combination with LPV/RTV. In addition, there is no efficacy data for using RBT to treat TB in children. Safety and efficacy exist for using RBT to treat Mycobacterium Avium Complex (MAC) infections.

The initial RBT dose of 5 mg/kg three times weekly is based on a review of available pediatric and adult data (see details below) and corresponds to the recommended dose of 150 mg/kg 3 times per week in adults. Because there is no direct benefit in the participating children, we opt for a conservative, safe approach and will not adjust the target RBT concentrations upwards, even if the dose recommendation for adults increases in the future (which could happen based on preliminary results presented by Boulanger et al) [29].

Data in Adults

• In adults, the recommended dose for RBT when given together with LPV/RTV is 150 mg 3 times per week, which is half the dose at half the frequency compared to administration of RBT without LPV/RTV [21, 28].
• Data in adults has demonstrated that the development of acquired rifamycin resistance is more likely in those with AUC 3.0 (1.9-4.5) [mean (95% CI)] compared to those with AUC 5.2 (4.6-5.8) [mean (95% CI)] [30].

Data in Children

• The USPHS/IDSA guidelines suggest RBT 5mg/kg QD in children for preventing MAC infection without concomitant protease inhibitors, and using half this dose when combining rifabutin with PIs.
• Dose for MAC treatment in children is 18.5 mg/kg (age ≤1 year) and 8.6mg/kg (age 2-10 years), with maximum of 300 mg [31].
• The minimal inhibitory concentration is higher for Mycobacterium Avium Complex (MAC), so treating TB based on MAC dosing should be sufficient.
• A RBT dose of 5mg/kg three times a week for two weeks will cause minimal toxicity in children as toxicity has been observed in populations that used much higher doses (mean 9.2mg/kg) and for a much longer duration (mean of 27 weeks) (Smith).
• Safety and efficacy data in treating MAC in children are sparse. The risk for corneal deposits is increased with prolonged administration of RBT [32], especially after 14 months of treatment, which is much longer than the standard 6 month treatment for TB.
• "There is a nonlinear relationship between clearance and size. The linear per kilogram model, based on adult values, underestimates clearance and, consequently, maintenance dose in children. An established scientific framework is supportive of an allometric-based model for weight with a coefficient of 0.75 for clearance and 1 for volume" [33]. This corresponds to approximately twice the clearance and consequently twice the per kilogram dose requirement in children of about 2-years old (range: 1-5 years) [34].

Interim Analyses to evaluate the need for dose adjustment An interim analysis of the PK data will be performed after either completion of study activities in 10 children or the occurrence of a SAE determined to be related to the RBT, whichever occurs first. A SAE will be defined as any grade 3 or 4 laboratory abnormality, development of uveitis, hospitalization or death. Every SAE will be reviewed by a committee consisting of three external reviewers: Dr. Mark Cotton and Dr. Brian Eley, infectious disease pediatricians in Cape Town, SA, and Dr. Ed Caparelli, a US based pharmacologist. If the committee decides that the SAE is unrelated to RBT, enrollment will continue until 10 children are enrolled or a second SAE occurs. If the SAE is judged to be related to RBT, then the interim analysis will be performed prior to completion of study activities in 10 children. Children in whom the study drug has been started will continue administration for all 6 doses, unless the committee decided that all drug administration needs to be halted. The NIH program officer (Carol Worrell) will be informed of each SAE and the decision made by the committee.

Based on the interim analysis of PK data, a decision will be made on the dose and dosing interval in subsequent children and the final sample size (determined based on variability observed in the first 10 children). The sample size will aim to have sufficient precision of the dose that reaches a geometic mean AUC (0-48 h) in children of 5-6 mg.h/L, approximating the AUC (0-48h) for rifabutin shown to be effective in adults [30] and approximating RBN AUC when used alone in a dose of 300 mg/d in adults [35].

PK Sampling and safety assessment PK sampling will consist of one intensive PK assessment for steady state plasma concentrations of RBT, 25-o-desacetyl RBT (desRBT, the most important active RBT metabolite) and LPV after 14 days of co-administration with sampling at 0, 2, 4, and 9 hours for LPV and 0, 2, 4, 9, 24 and 48 hrs for RBT and desRBT. To perform LPV, RBT, and desRBT assays plasma concentrations, a total of 1.5 ml whole blood will be collected.

Safety assessment (including clinical exam, liver function test (ALT), and full blood count) will be performed at screening, at the first scheduled visit (within 2 to 4 weeks of last RBT administration), and at any time the child presents with signs of symptoms of possible RBT associated toxicity.

Concomitant Medications The following medications are not allowed during the study: any antiretrovirals other than LPV/RTV and NRTIs, Rifampin, Phenytoin, Phenobarbital, Carbamazepine, Fluconazole, Ketoconazole, Itraconazole, Steroids, any other inducers of inhibitors of the cytochrome P450 and PGP enzyme system.(see appendix) Any additional medications required during the study period should be approved by a member of the study team.

LABORATORY EVALUATIONS Processing of peripheral blood samples Whole blood samples of 1.5 ml will be drawn in lithium heparin tubes, centrifuged at 4degrees C at 2600rpm for 10min; plasma removed; shipped and stored at -70 degrees C for future analysis in the laboratory of Dr Maartens, University of Cape Town.

LPV concentrations will be quantified by a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) method using a modification of the method by Chi J et al [36].The calibration curve is linear over the range 0.05-20 mg/L. Any sample with LPV results above 20 mg/L will be diluted with drug-free plasma and reanalyzed. The lower limit of quantification (LLOQ) of the LPV assay is 0.05 mg/L. Any samples with LPV concentration below the LLOQ will be reported as <0.05 mg/L and treated as 0 mg/L in the analysis. LPV assay accuracy ranged from 94.3 to 103.0%. The intra-day and inter-day precisions ranged from 0.14 to 4.72% and from 1.61 to 4.22%, respectively.

Assays for RBT and its active metabolite 25-o-desacetyl RBT have been developed on the LC-MS/MS, The validated concentration range for both RBT and 25-o-desacetyl RBT is 5-500 ng/mL (=0.005 to 0.5 mg/L).

Screening and Safety Laboratory Assessments Laboratory assessments performed at screening and required for toxicity reasons during the study will be conducted at the NHLS laboratory.

CLINICAL MANAGEMENT ISSUES

Toxicity Subjects may experience toxicities and side effects such as uveitis, arthritis, liver toxicity, or neutropenia. All toxicities are rare at the dose used in the study, reversible upon interruption of RBT, and have only been described in children receiving RBT for several months.

To maximally limit the potential risk of toxicity occurring, we will:

• carefully screen children for pre-existing condition that could increase the risk of toxicity (see eligibility criteria)
• Perform careful clinical and laboratory safety monitoring and use the DAIDS/IMPAACT grading system to grade toxicities.
• Have all SAE (death, hospitalization, grade 3 or 4 AE) reviewed by an independent safety committee (see section 5.1.2). Enrollment will be halted until the committee decides whether the SAE is related to RBT or not.

CRITERIA FOR DISCONTINUATION

- Premature Study Discontinuation Any child that develops a rash or an SAE (grade 3 or 4 laboratory abnormality, development of uveitis, or hospitalization) will be discontinued from the study.

STATISTICAL CONSIDERATIONS

Endpoints The primary endpoint of the study is the dose of RBT with a resultant AUC of 4.5-6.0 mcg/mL when used in combination with LPV/RTV in children age < 3 years.

Stratification If sample size and patient characteristics allow, analysis will be stratified by age and nutritional status (weight for age z score ≤ or > minus 3).

Sample Size and Accrual We estimate that a sample size of 10 children will provide information on variability of the AUC parameter in children, sufficient to calculate a final sample size at the time of the interim analysis. We estimate that a final sample size of 30-40 will allow a sufficient number of samples in order to perform PK modeling (210 data time points) and stratify analysis by age and nutritional status.

Analyses LPV/RTV and RBT concentration data will be analyzed using standard WinNonLin Pro software to determine PK parameters. Drug exposure will be assessed using AUC0-24hr and AUC0-48hr for RBT and desRBT, and AUC0-12hr for LPV. We will also assess C0hr, Cmax, Tmax, and oral clearance (CL/F) for RBT, desRBT and LPV. The AUC will be determined using the trapezoidal rule. To analyze factors associated with the RBT serum levels, we will perform analysis of covariance (ANCOVA), with rifabutin AUC0-24hr adjusted for drug dose as the dependent variable and demographic, clinical, anthropometric, nutritional and laboratory characteristics as independent variables. In the final analysis population pharmacokinetic models for RBT and desRBT will be generated using nonlinear mixed effects methods. These models will be used to simulate the RBT doses required for pharmacokinetic targets, thus allowing dose recommendations in subpopulations and, simulation of changing RBT exposure targets. The data will also be used to validate population RBT models built on data from other sources

HUMAN SUBJECTS Institutional Review Board (IRB) Review, Recruitment and Informed Consent The study will be reviewed by the Institutional Review Board of the University of North Carolina at Chapel Hill, the University of the Witwatersrand and the University of Cape Town. Approval of the study will be needed before any study activities are started.

The study will also be submitted to the South African Medicine Control Council for review and approval.

Potential subjects will be identified during routine clinic visits. The health care worker in charge of the care of the child will briefly explain the study to parent or legal guardians of children meeting eligibility requirements. Caretakers that are interested will be referred to the study nurse for the informed consent process. The informed consent will describe the purpose of the study, the procedures to be followed and the risks and benefits of participation. In the event the study subject's guardian cannot read, this consent will be performed verbally in the subject's native language. All informed consent procedures will be performed by trained study personnel in a private exam room to ensure privacy. A copy of the consent from will be given to the subject.

General Design Issues The study of novel drug regimens in children is important but poses several ethical challenges. The study team (and NIH reviewers) assessed several approaches to the study of RBT PK in HIV infected children in need for concomitant ART and TB treatment.

Children could have a direct benefit of receiving RBT if in need of TB treatment, i.e HIV-infected children diagnosed with active TB. Enrollment of these children in a RBT dose finding study however raises the possibility of substandard treatment for TB, a risk which is not defendable.

To reduce the risk of exposure to substandard TB treatment, one could perform real time therapeutic drug monitoring. Unfortunately, "real time" therapeutic drug monitoring still has a turn-around time of at least one week, which still poses the an unacceptable risk of substandard treatment for TB in these children during that time period. Seen that TB has substantial risk of death in HIV infected children, this risk is unacceptable.

Following several rounds of discussion, the group of experts decided that RBT PK should only be studied in "healthy children", i.e. children free of TB. Because of the difficulty in excluding active TB in children (obtaining a respiratory specimen in young children is difficult and even if successful, the bacteriological yield is low due to the paucibacillary nature of TB in children), it was decided that the optimal characteristics of eligible children are those who recently successfully completed TB treatment (see section 4.1).

Realizing that HIV infected children are potentially more vulnerable compared to HIV uninfected children, one could suggest including HIV negative children who recently completed TB treatment. The disadvantages to this approach is that one exposes children to two drugs they do not need (no direct benefit from either drug) compared to one drug (in the proposed study, only children receiving LPV/RTV are eligible). Furthermore, the drug absorption and metabolism in HIV infected children may differ that of HIV uninfected children, a phenomenon that has been observed for TB drugs in HIV infected vs. HIV negative adults [37-40]. For these reasons, we selected to include HIV infection as an eligibility criterion.

Subject Confidentiality Data collected for this study will be entered into a confidential database. The patients will not be assigned a study number, all data will be stored and analyzed by reference to the study number rather than to the patient name. Study specific blood samples will also be labelled according to the study number in order to preserve anonymity. It will not be possible to determine the identity of the patients from the information maintained in the database. The database will be password protected with access restricted to certain study personnel (principal investigators A.V.R., T.M., H.M., G.M., study nurse, and relevant consultants. All documents specific to this research linking patient personal information to study ID code numbers will be kept securely in SA. Only the principal investigators, study manager and study nurse will have access to these files.

PUBLICATION OF RESEARCH AND RESEARCH FINDINGS Research findings will be disseminated at relevant local (South African) and international meetings, and published in local and/or international medical journals, as appropriate.

SELECTION AND ENROLLMENT OF SUBJECTS

Inclusion Criteria:

• Children with confirmed HIV infection. Confirmation can be by two rapid tests (children age > 18 months) or virologic test (children < 18 months), and detectable viral load prior to starting ARVs.
• Age ≤ 5 years old - rationale: changes in body composition and maturity of metabolizing enzymes and organs result in age-related differences in drug clearance, especially between children ≤ 5 years and children > 5 years of age.
• Receiving an ART regimen containing 2 NRTIs + LPV/RTV at standard dose
• Successfully completed TB treatment in the past 2 to 6 weeks. Successful completion of treatment will be defined as children with good clinical response (resolution of TB symptoms) to treatment.

Rationale:

1. RBT has not yet been approved for treatment of TB in children. Participating children can therefore not be in need treatment for TB as this may lead to substandard treatment.
2. RBT monotherapy in the presence of Mycobacterium tuberculosis can lead to the development of resistance. Excluding active TB is difficult in children, especially those that are HIV co-infected. Children who have just successfully completed a treatment for TB can be assumed to be free of Mycobacterium tuberculosis.
3. A minimum of two weeks is needed between RIF and RBT administration to ensure wash-out of any enzyme inducing effects of RIF.

Exclusion Criteria

• History of symptomatic clinical hepatitis during TB treatment
• Abnormal liver function defined as ALT > 2.5 times the normal upper limit (corresponding to the US National Institute of Health Division of AIDS scale grade 2)
• Abnormal bilirubin defined as > 1.5 UNL (≥ DAIDS grade 2)
• Abnormal serum creatinine defined as >1.1 x ULN
• Anemia defined as hemoglobin < 8gm/dL
• Neutropenia defined as < 1.0 x 109/L(corresponding to grade 2)
• Abnormal platelets defined as <125 x 1012/L
• Pre-existing eye conditions
• Any condition that the clinician feels would predispose the child to toxicity
• Children required to take any drug known or predicted to interact with rifabutin (see appendix)
• Children who do not meet inclusion criteria and/or whose parents refuse consent.