Model‐Informed Once‐Daily Dosing Strategy for Bedaquiline and Delamanid in Children, Adolescents and Adults with Tuberculosis

Abstract

The complexity of the currently registered dosing schedules for bedaquiline and delamanid is a barrier to uptake in drug‐resistant tuberculosis treatment across all ages. A simpler once‐daily dosing schedule is critical to ensure patient‐friendly regimens with good adherence. We assessed expected drug exposures with proposed once‐daily doses for adults and compared novel model‐informed once‐daily dosing strategies for children with current World Health Organization (WHO) recommended dosing. A reference individual and virtual pediatric population were generated to simulate exposures in adults and children, respectively. Published population models characterizing the exposures of bedaquiline and its metabolite M2, delamanid, and its metabolite DM‐6705 were utilized. During simulation, child growth during treatment along with several CYP3A4 ontogeny profiles was accounted for. Exposures in children were compared with simulated adult targets to assess the expected treatment efficacy and safety. In adults, the proposed bedaquiline once‐daily dosing (400 mg daily for 2 weeks followed by 100 mg daily for 22 weeks) yielded 14% higher exposures of bedaquiline and M2 compared to the labeled dosing scheme at 24 weeks; for delamanid and DM‐6705, the suggested 300 mg daily dose provided 13% lower exposures at steady state. For children, the cumulative proportions of exposures of both drugs showed < 5% difference between WHO‐recommended and proposed once‐daily dosing. This study demonstrated the use of model‐informed approaches to propose rational and simpler regimens for bedaquiline and delamanid in adults and children. The new once‐daily dosing strategies will be tested in the PARADIGM4TB and IMPAACT 2020 trials in adults and children, respectively.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Current dosing schedules for bedaquiline and delamanid are complex and difficult to use in combination with other anti‐tuberculosis medications. A simpler once‐daily dosing schedule is critical to enable patient‐friendly regimens and improve adherence; however, there are limited data available.

• WHAT QUESTION DID THIS STUDY ADDRESS?

Whether the novel model‐informed once‐daily dosing regimens for bedaquiline and delamanid could achieve similar exposures associated with efficacy and safety compared to the labeled or World Health Organization (WHO)‐recommended dosing regimens in adults, adolescents, and children.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

Our simulation results demonstrated that newly proposed once‐daily dosing strategies for bedaquiline (400 mg daily for the first 2 weeks, followed by 100 mg daily for 22 weeks) and delamanid (300 mg daily for 24 weeks) showed similar pharmacokinetic profiles in adults compared to the labeled dosing scheme. Exposures for both drugs and their metabolites in children were similar between the WHO‐recommended and model‐informed novel daily dosing regimens.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

The model‐informed approaches allow for predicting exposure profiles among adults and a variety of age‐weight band groups in children under various scenarios. This can help justify the proposed dose selection.

Tuberculosis (TB) is a life‐threatening infectious disease and remains a serious public health challenge. In 2022, approximately 10.6 million people were newly diagnosed with TB, and 410,000 people developed rifampicin‐resistant or multidrug‐resistant TB (MDR‐TB). ¹ The introduction of novel drugs, e.g., bedaquiline, has revolutionized the treatment of drug‐resistant TB across all ages, while delamanid provides an alternative treatment option for patients. ² Current regulatory approved dosing for bedaquiline in adults is 400 mg daily for a 2‐week loading phase, followed by 200 mg thrice‐weekly in the 22‐week continuous phase ³ ; for delamanid it is 100 mg twice‐daily dosing for 24 weeks. ⁴ The efficacy and safety of bedaquiline and delamanid in the regulatory approved dosing are well‐established in adults. ⁵ , ⁶ , ⁷ , ⁸ Nonetheless, the complexity of the currently registered dosing schedules for both drugs is a barrier to uptake. Simplification to once‐daily dosing is critical to enable patient‐friendly regimens as well as future fixed‐dose combinations and ultimately improve adherence and patient‐centered outcomes. ⁹ Currently there are limited data for both drugs in once‐daily dosing regimens. In this work we listed the key considerations for designing such regimens and proposed novel model‐informed once‐daily dosing strategies for bedaquiline and delamanid in adults, adolescents, and children.

For bedaquiline, it is important to have high exposure in the first 2 weeks to achieve the greatest early bactericidal activity (EBA) ¹⁰ while maintaining low concentration of its primary metabolite M2 over a 24‐week treatment period to avoid QT prolongation. ¹¹ , ¹² In the ZeNix trial, once‐daily dosing of bedaquiline (200 mg daily for 8 weeks followed by 100 mg daily for 16 weeks) was tested in the bedaquiline‐pretomanid‐linezolid (BPaL) regimen, ¹³ and no significant safety concern due to bedaquiline was identified. This once‐daily bedaquiline dose has been suggested for use in adults in the World Health Organization (WHO) recommendations in 2022. ¹⁴ Although this daily dosing approach provided comparable daily average exposures to the thrice‐weekly dosing based on a pharmacokinetic (PK) simulation study, ¹⁵ it lacks the intensive loading phase in the initial 14 days, which is expected to result in a lower EBA, a potential concern. ¹⁰

Delamanid exposure has been found to be associated with efficacy, ¹⁶ and concentration of its metabolite DM‐6705 is relevant to QT prolongation. ¹² In addition, nonlinear bioavailability with increased delamanid dose has been identified. ¹⁷ , ¹⁸ , ¹⁹ This means that a higher total daily dose will need to be administered when given once instead of twice daily to achieve comparable exposures. Once‐daily dosing of delamanid has been investigated in a 14‐day EBA study in adults ¹⁸ and in the PHOENIx trial (AIDS clinical trials group ACTG5300/IMPAACT2003, ClinicalTrial.gov identifier: NCT03568383). ²⁰ The maximum concentration (C _max) and daily area under the concentration‐time curve (AUC_0‐24h) of delamanid at the dose of 300 mg daily were higher than 200 mg daily but lower than the approved 100 mg twice‐daily dose. ⁷ , ¹⁸

Approximately 25,000–32,000 MDR‐TB cases occur in children < 15 years of age annually. ²¹ Given the complexity and practical considerations of undertaking clinical trials in children, pediatric data on bedaquiline and delamanid are limited. Furthermore, simple allometric scaling on the basis of body weight from adults to “small adults” often results in incorrect predictions of pediatric dose for children under 2 years of age, due to the complexities of their physiology, organ function, as well as maturation of enzymes. ²² In 2022, WHO updated the guideline for management of drug‐resistant TB in children and adolescents and now recommends bedaquiline at a 2‐week daily loading dose plus a 22‐week thrice‐weekly dosing and twice‐daily delamanid for all children > 3 kg. ²³ The majority of the evidence for the pediatric dosing of bedaquiline and delamanid came from modeling and simulations and emerging exposure and safety data from pediatric trials, using a similar approach as presented in this work. ²⁴ With quantitative modeling and simulation methods, existing knowledge of drug properties and population characteristics can be best leveraged to facilitate decision‐making in dose selection and optimization. ²⁵ , ²⁶ Growth and maturation of enzymes in children, especially neonates and young infants, can be accounted for during model construction and simulation under different dosing scenarios. ²⁷ , ²⁸ In this study, we predicted the drug exposures with proposed once‐daily doses of bedaquiline and delamanid for adults and children. We then compared the novel model‐informed once‐daily dosing regimens with the currently used dosing approaches. This work supports ongoing and future clinical trials evaluating once‐daily dosing strategies for bedaquiline and delamanid in adults in short and patient‐friendly regimens (e.g., PARADIGM4TB, ClinicalTrial.gov identifier: NCT06114628) and in children (e.g., IMPAACT 2020).

METHODS

Evaluated dosing regimens

In this work, we evaluated the following three regimens for bedaquiline in adults: the registered dose (400 mg once daily for 2 weeks, followed by 200 mg thrice weekly for 22 weeks), ³ the dose used in the ZeNix trial (200 mg once daily for 8 weeks, followed by 100 mg once daily for 16 weeks), ¹³ and the new suggested once‐daily dose (400 mg once daily for 2 weeks, followed by 100 mg once daily for 22 weeks). For delamanid in adults, two regimens were assessed: the 100 mg twice daily (registered dose) ⁴ and 300 mg once daily (suggested dose) dosing for 24 weeks.

For children, two joint age‐ and weight‐tiered dosing strategies (i.e., WHO‐recommended dosing ²³ and novel once‐daily dosing suggested by our study) based on commercially available tablets and the weight bands used in the current WHO MDR‐TB treatment guidance were evaluated. Recently proposed harmonized weight bands across the therapeutic areas ²⁹ were also assessed.

Virtual population

For deriving the typical concentration‐time profiles for both drugs in adults, a typical individual from the published population PK model (Svensson et al. ³⁰ ) aged 32 years, of non‐Black race, weighing 56.6 kg, and having an albumin level of 3.65 g/dL at baseline was used. Simulation of virtual adult population to explore exposure variability is detailed in the Supplementary Material . To generate a realistic virtual pediatric population representative of children with TB, 40,000 pediatric patients were simulated with age uniformly distributed from 0 to < 15 years with a 50%/50% gender split and 40% of Black race. The weight‐per‐age distribution was based on WHO growth curves (0–10 years) ³¹ and data from the National Health and Nutrition Examination Survey (NHANES) (11–15 years), ³² adjusted by a TB disease factor. ³³ Detailed methodology has been described elsewhere. ³³ Preterm infants or infants with weight below 3 kg were not evaluated here. Children with extremely high weight (> 80 kg, i.e., the 95th percentile of a 15‐year‐old child) were excluded.

During simulation, both drugs were assumed to be given with food. Additionally, child growth was taken into account due to the long duration of TB treatment. For each individual, age and weight were checked at the start of each month, and individuals would be assigned to the correspondent weight‐age‐tiering groups. Dose would thereafter be adjusted accordingly. Simulation periods for bedaquiline and delamanid, together with their metabolites, were set as 24 and 8 weeks, respectively, given prior knowledge of their half‐lives. Given the complexity of growth, physiological, and metabolic changes for young infants and neonates, evaluation of dose selection in children under 1 year old was of particular interest. Therefore, patients were stratified in four groups according to their age at the start of simulation (0 to < 3 months, 3 to < 6 months, 6 to < 12 months, ≥ 1 year at start) for dose evaluation.

Population PK models for simulation

Published population models characterizing the exposures of bedaquiline and its metabolite M2 (Svensson et al. ³⁰ ) as well as delamanid and its metabolite DM‐6705 (Tanneau et al. ³⁴ ) were selected to simulate typical adult exposures. In summary, the Svensson model, built on 335 patients from two phase II trials, jointly described bedaquiline and M2 exposures with incorporation of semi‐physiological changes of weight and albumin levels over time. Race and age were identified as significant covariates on clearance. For delamanid and DM‐6705 exposures described by the Tanneau model using data from 52 participants, the adherence effect included after week 8 and time of intake (morning/evening) had a significant impact on bioavailability. Albumin concentrations did not affect delamanid and DM‐6705 PK. Some changes were adapted while using the Tanneau adult model: (i) The adherence effect on bioavailability identified in the model was not included; (ii) daily average bioavailability without morning/evening dose effects was used in once‐daily dosing simulation settings; (iii) a 42% decrease in bioavailability while receiving a dose > 200 mg was implemented given the results from previous population PK analysis. ¹⁷

For simulation of bedaquiline and M2 exposures in children, the Svensson adult model ³⁰ served as a basis and was adapted to the pediatric population with the following assumptions: (i) The standard value commonly used in the scaling of clearances is 0.75, supported by current pediatric studies, ²⁴ which was implemented instead of 0.18, as described in the adult model; (ii) the effect of albumin concentrations on clearances was not accounted for, as children with TB were found to have a relatively normal range of albumin levels; and (iii) the maturation function of cytochrome P450 3A4 enzyme (CYP3A4) proposed by Johnson et al. ³⁵ was adopted. For delamanid and DM‐6705, the pediatric model from Sasaki et al. ¹⁹ based on data from 37 children in two pediatric delamanid trials was utilized. The identified covariates included (i) allometric scaling with weight on clearance (exponent 0.75) and volume of distribution (exponent 1); (ii) dose effect on bioavailability; (iii) formulation effect on bioavailability and mean absorption time, and (iv) age effect on fraction absorbed and bioavailability. Notably, the effect of age on bioavailability identified was uncertain for those aged below 1 year old, given that only one participant < 1 year old was included in the analysis. Hence, we simulated two scenarios in children < 1 year of age: one with linearly decreasing bioavailability with age and the other with the same bioavailability as 1‐year‐old children. Detailed model structure and identified covariates are described in the original publications. ¹⁹ , ³⁰ The population PK models used for simulation in children were supported by data from several pediatric studies. ²⁴

Given that CYP3A4 plays a pivotal role in bedaquiline, M2, and DM‐6705 metabolism, ³ , ⁴ , ³⁶ several CYP3A4 ontogeny profiles (Johnson, Salem, and Upreti) ³⁵ , ³⁷ , ³⁸ were assessed in a sensitivity analysis for the pediatric population. Details of CYP3A4 maturation functions are provided in the Supplementary Material and Figure S1 .

Target exposure metrics

The predicted exposures in children under the WHO‐recommended and model‐informed once‐daily dosing regimens were compared with model‐derived adult exposure matching targets. PK target metrics for bedaquiline and delamanid at varying time points were selected. Weekly AUC (AUC_0‐168h) of bedaquiline and daily AUC (AUC_0‐24h) of delamanid have been found to drive the efficacy. ¹⁶ , ³⁹ For safety evaluation, maximum concentration (C _max) for M2 and DM‐6705 was selected based on the identified correlation between the plasma concentration and QTc prolongation. ¹¹ , ¹² Given that the bedaquiline dose was identical in the initial 14 days under the current WHO‐recommended and newly proposed once‐daily regimens, PK metrics at day 14 for bedaquiline and M2 were not of interest in this work. Eight and 24 weeks were chosen to evaluate bedaquiline dosing regimens in the maintenance phase. For delamanid and DM‐6705, we assumed that a steady state was reached by day 14 and week 8, respectively. ⁷ , ⁴⁰ As a result, AUC_0‐24h of delamanid at day 14, C _max of DM‐6705 at day 14, and week 8 were assessed. All PK‐matching targets were derived under the registered dosing regimens from the aforementioned adult models, as exposure‐response relationships were assumed to be similar in children and adults. ²⁶ , ⁴¹ The 5th and 95th percentiles of simulated PK targets were used to evaluate efficacy. Nevertheless, only the upper limit (i.e., 95th percentile) was of our interest in metabolites while considering safety. Descriptions of deriving target exposures are detailed in the Supplementary Material .

Software

A Virtual population of children was generated in NONMEM nonlinear mixed effects modeling software (version 7.5.1). ⁴² Adult virtual population, target simulation, data management, post‐processing, and graphical analyses were performed in R (version 4.3.1 and higher). All simulations were performed by using the mrgsolve package (version 1.3.0) in R. ⁴³

RESULTS

Simulated typical exposures of bedaquiline and delamanid in adults

Simulated typical average concentration profiles of bedaquiline and delamanid, together with maximum concentration profiles of their metabolites in adults under evaluated regimens, are presented in Figure 1 . Both the registered dosing and novel model‐informed once‐daily dosing strategies for bedaquiline yielded the peak of concentration at day 14, whereas the highest exposure in the ZeNix dosing scheme occurred at week 8, i.e., the end of the period given in a higher dose. Compared to the registered regimen, exposures of bedaquiline and M2 were 14% higher in the novel model‐informed once‐daily dosing by the end of 24‐week treatment. Under ZeNix dosing, the cumulative AUC of bedaquiline was 50% lower than the registered and suggested once‐daily dosing at day 14 but reached similar levels at 8 weeks. At 24 weeks, the ZeNix dosing approach yielded approximately 20% higher exposures than the registered dosing. For delamanid and DM‐6705, the proposed 300‐mg once‐daily dose provided 13% lower exposures at steady‐state. Predicted typical exposures of both drugs and their metabolites are summarized in Tables S1 and S2 . Full concentration‐time profiles for a typical adult individual are provided in Figure S2 . Predicted exposures of both drugs in a representative adult population are displayed in Figure S3 .

Figure 1.

Typical pharmacokinetic profiles of (a) bedaquiline, (b) M2, (c) delamanid, and (d) DM‐6705 over time for a reference individual (32‐year‐old, non‐Black race, 56.6 kg, and 3.65 g/dL of albumin level at baseline). Mean concentration over the dosing interval is used as the exposure metric for bedaquiline, mean daily concentration for delamanid, and maximum concentration for M2 as well as DM‐6705. Line types represent investigated dosing regimens: registered (solid line), suggested once‐daily (dashed line), and ZeNix (dotted line, only for bedaquiline and M2) dosing strategies. BID, twice daily; QD, once daily; TIW, thrice weekly.

Simulation of exposures of bedaquiline and delamanid in children

Population

A total of 39,993 virtual children ranging from 0 to < 15 years and weighing between 3 and 80 kg were used in the simulations. The median (range) of weight in children < 1 year of age stratified in 0 to < 3 months, 3 to < 6 months, and 6 to < 12 months age groups was 5.0 (3.2–7.7) kg, 6.1 (4.0–9.1) kg, and 7.4 (4.3–11) kg, respectively. A full distribution in age‐weight tiers of the simulated pediatric population is demonstrated in Table S3 .

Proposed novel once‐daily dosing

For bedaquiline, we kept the high loading dose in the early 14‐day phase and halved the registered dose in the continuous phase. For delamanid, the dosing schedules were proposed based on the knowledge of its non‐proportional dose‐dependent bioavailability. Table 1 displays the WHO‐recommended and proposed once‐daily dosing for bedaquiline and delamanid in children < 15 years of age by WHO MDR‐TB weight bands. Details of the once‐daily dosing schemes utilizing harmonized weight bands are presented in Tables S4 and S5 .

Table 1.

World Health Organization (WHO)‐recommended dosing and proposed once‐daily dosing regimens of bedaquiline and delamanid for children and adolescents aged below 15 years old with multidrug‐resistant tuberculosis, applying the age‐ and weight‐based bands currently used in WHO dosing guidelines for multidrug‐resistant tuberculosis

Bedaquiline				Delamanid
Age	Weight	WHO dosinga	Daily dosinga	Age	Weight	WHO dosing	Daily dosing
0 to < 3 months	All	30 mg QD/10 mg TIW	30 mg QD/5 mg QD	0 to < 3 months	All	25 mg QD DT	25 mg QD DT
3 to < 6 months	All	60 mg QD/20 mg TIW	60 mg QD/10 mg QD	3 to < 6 months	3 to < 5 kg 5 to < 10 kg	25 mg QD DT 25 mg BID DT	37.5 mg QD DT b 37.5 mg QD DT b
≥ 6 months	3 to < 7 kg	60 mg QD/ 20 mg TIW	60 mg QD/ 10 mg QD	≥ 6 months	3 to < 5 kg	25 mg QD DT	37.5 mg QD DT
	7 to < 10 kg	80 mg QD/40 mg TIW	80 mg QD/20 mg QD		5 to < 10 kg	25 mg BID DT	37.5 mg QD DT
	10 to < 16 kg	120 mg QD/60 mg TIW	120 mg QD/30 mg QD		10 to < 16 kg	25 mg BID DT	50 mg QD DT
	16 to < 30 kg	200 mg QD/100 mg TIW	200 mg QD/50 mg QD		16 to < 30 kg	50 mg morning DT, 25 mg evening DT	100 mg QD DT
	≥ 30 kg	400 mg QD/200 mg TIW	400 mg QD/100 mg QD		30 to < 46 kg	50 mg BID	150 mg QD
					≥ 46 kg	100 mg BID	300 mg QD

^a Bedaquiline dosing is indicated as a loading dose/maintenance dose. For example, “30 mg QD/10 mg TIW” indicates 30 mg daily for 2 weeks followed by 10 mg thrice weekly for 22 weeks.

^b All children aged 3 to < 6 months, regardless of weight, received a 37.5 mg daily dose of delamanid under the proposed once‐daily dosing strategy. BID, twice daily; DT, dispersible tablet; QD, once daily; TIW, thrice weekly; WHO, World Health Organization.

Simulation of bedaquiline exposures

Figure 2 depicts the cumulative proportions of the population with varying bedaquiline and M2 exposures using WHO MDR‐TB weight bands. Predicted drug exposures in cumulative curves were presented as the exposures using Johnson ontogeny as a representative CYP3A4 maturation model. The cumulative curves were in principle similar between the two regimens, as demonstrated by < 5% difference across age strata. Overall, higher exposures of bedaquiline (9.1–15%) and M2 (11–22%) under the proposed once‐daily dosing were observed compared to current WHO‐recommended dosing schemes. Under the proposed once‐daily dosing schedule, 84–90% of patients were predicted to be within the desired AUC_0‐168h target range of bedaquiline, and over 90% of the population were below the 95th percentile of M2 C_max target. Figures S4 and S5 showed that acceptable bedaquiline and M2 exposures were reached by once‐daily dosing under harmonized weight bands.

Figure 2.

Cumulative proportions of the pediatric population of predicted (a) bedaquiline AUC_0‐168h and (b) M2 C _max at week 8 and week 24 using the Johnson CYP3A4 ontogeny. Solid curves represent the pediatric population. Dashed curves represent adult population under the regulatory approved dosing regimen (400 mg once daily for the initial 2 weeks, followed by 200 mg thrice weekly for 22 weeks). Vertical lines represent the 5th and 95th percentiles of bedaquiline target exposures (95th percentile for M2) derived from the adult population. Shaded areas and the correspondent numbers in percentages represent the proportion of the population under the lower limit (5th percentile) or above the upper limit (95th percentile) of bedaquiline or M2 target exposures. AUC_0‐168h, area under the concentration‐time curve over 168 hours; C _max, maximum concentration.

Figure 3 demonstrates simulated bedaquiline and M2 exposures in children aged < 1 year using WHO MDR‐TB weight bands under different assumptions of CYP3A4 maturation profiles (Figure S6 for children ≥ 1 year old). At 24 weeks, more than 89% of children aged < 1 year were within the target range of bedaquiline and M2 while using the Johnson and Salem ontogenies, whereas 34% of children < 1 year of age were below the 5th percentile of bedaquiline weekly AUC target by using the Upreti ontogeny. Compared to the Johnson model, the Salem model resulted in approximately 30% higher bedaquiline and M2 exposures, and the Upreti model predicted approximately 50% lower in the youngest group (0 to < 3 months at the start of the simulation) by the end of the 24‐week treatment.

Figure 3.

Predicted (a) bedaquiline AUC_0‐168h and (b) M2 C _max in the pediatric population aged below 1 year old at week 8 and week 24. Horizontal lines indicate the simulated adult target range (5th and 95th percentiles in dashed lines and median in solid lines) at a regulatory‐approved bedaquiline dose (400 mg QD for the initial 2 weeks, followed by 200 mg TIW for 22 weeks). Patterns of boxplots represent different age‐specific CYP3A4 maturation functions. AUC_0‐168h, area under the concentration‐time curve over 168 hours; C _max, maximum concentration; QD, once daily; TIW, thrice weekly.

Simulation of delamanid exposures

The cumulative proportions of the population with varying delamanid and its metabolite DM‐6705 exposures using WHO MDR‐TB weight bands at selected timepoints are shown in Figure 4 . Individual exposures presented in cumulative curves were simulated using the Sasaki pediatric model without CYP3A4 ontogeny functions on DM‐6705 metabolism. The WHO‐recommended or suggested once‐daily dosing regimens were the same in those < 3 months, leading to identical exposures of delamanid and DM‐6705. More than 96% of patients receiving a once‐daily dose were within desired target ranges for both delamanid and DM‐6705 across age groups. The once‐daily dosing schemes under harmonized weight bands resulted in acceptable exposures of delamanid and DM‐6705 (Figures S7 and S8 ).

Figure 4.

Cumulative proportions of the pediatric population of predicted (a) delamanid AUC_0‐24h at day 14 and (b) DM‐6705 C _max at day 14 as well as week 8. Solid curves represent the pediatric population. Dashed curves represent adult populations under a regulatory approved dosing regimen (100 mg twice daily for 24 weeks). Vertical lines represent the 5th and 95th percentiles of delamanid target exposures (95th percentile for DM‐6705) derived from the adult population. Shaded areas and the correspondent numbers in percentages represent the proportion of the population under the lower limit (5th percentile) or above the upper limit (95th percentile) of delamanid or DM‐6705 target exposures. AUC_0‐24h, area under the concentration‐time curve over 24 hours; C _max, maximum concentration.

Figure 5 illustrates that among three CYP3A4 maturation functions, the highest exposures of DM‐6705 were predicted with the Salem function and the lowest with the Upreti function. In children < 3 months of age, the constant bioavailability model (no further decrease of bioavailability in children < 1 year old) predicted approximately 22% higher median exposures of delamanid and DM‐6705 at steady‐state compared to the linear model. The simulated exposures under different assumptions of CYP3A4 ontogeny functions in children ≥ 1 year are presented in Figure S9.

Figure 5.

Predicted (a) delamanid AUC_0‐24h and (b) DM‐6705 C _max in the pediatric population aged below 1 year old at day 14 (for delamanid and DM‐6705) and week 8 (for DM‐6705). Horizontal lines indicate the simulated adult target range (5th and 95th percentiles in dashed lines and median in solid lines) at a regulatory‐approved delamanid dose (100 mg BID for 24 weeks). Line colors of boxplots represent different assumptions on how age impacts bioavailability under 1 year of age. Patterns of boxplots represent different age‐specific CYP3A4 maturation functions. AUC_0‐24h, area under the concentration‐time curve over 24 hours; BID, twice daily; C _max, maximum concentration; QD, once daily.

DISCUSSION

In this work, we simulated typical adult exposures of bedaquiline and delamanid, including their metabolites, under the regulatory approved dosing and trial‐tested once‐daily dosing (for bedaquiline) in bedaquiline‐containing regimens like BPaLM/BPaL and novel model‐informed once‐daily dosing schemes. For bedaquiline, the high loading dose strategies in the registered and suggested model‐informed daily dosing yielded a rapid increase of exposure in the initial 14 days. This is thought to be advantageous as bedaquiline demonstrated a dose–response relationship in EBA studies. ¹⁰ Predicted exposures of bedaquiline and M2 under the proposed 100 mg once‐daily dosing in the maintenance period were slightly higher than those under the registered 200 mg thrice‐weekly dosing scheme. The expected exposures (bedaquiline AUC_0‐168 and M2 C _max) under the ZeNix dosing exceeded the exposures with the labeled dosing after 4 weeks of treatment start; nevertheless, modest bedaquiline‐related adverse effects were identified in the ZeNix trial. ¹³ Moreover, the predicted concentration of M2 is expected to increase QT interval by < 10 ms given the identified relationship between M2 exposures and QT prolongation. ¹¹ , ¹² The newly proposed once‐daily dosing strategy for bedaquiline is therefore assumed to be efficacious and safe for use. Our simulated results of delamanid in adults demonstrated a lower exposure in the proposed 300‐mg daily dose relative to the approved dosing strategy, which is consistent with the previous literature. ¹⁸ Despite this, the 14‐day EBA trial revealed significant bactericidal activity at a 300 mg once‐daily dose. ¹⁸ Prior exposure‐response analysis of delamanid also proposed a possible option of 300 mg once‐daily dosing, given the expected similar responses and lower risk of QT prolongation compared to the labeled 100 mg twice‐daily dosing. ¹⁶

For children, both WHO‐recommended and proposed once‐daily dosing schemes provided similar cumulative curves with varying exposures among bedaquiline, delamanid, and their metabolites, irrespective of age group categories. Despite systemic higher exposures of bedaquiline and the M2 in a once‐daily dosing regimen, our simulation showed that only up to 9.9% of patients were above M2 C _max target at the end of the treatment period under the once‐daily dose (5.9% under WHO‐recommended dose), indicating that safety is not expected to be a concern under once‐daily bedaquiline treatment in children. Notably, the WHO‐recommended bedaquiline doses are considerably lower than those evaluated in the registrational IMPAACT P1108 trial (ClinicalTrial.gov identifier: NCT02906007), where bedaquiline was safe and well‐tolerated. ²⁴ Patients under once‐daily dose of delamanid had lower exposures compared to the WHO‐recommended twice‐daily dose. However, these still closely approximated median adult exposures, and 96% of patients were in the acceptable delamanid exposure ranges in both regimens. Our simulation indicated that those receiving delamanid under a once‐daily dose had lower C _max of DM‐6705. Hence it is reasonable to conclude that the newly proposed once‐daily dosing strategy of delamanid for children may offer a better safety profile without sacrificing efficacy while offering a more patient‐friendly dosing regimen.

Different assumed ontogeny profiles could change predicted drug clearance dramatically. The expression levels of CYP3A4 in the Johnson and Salem models were lower than in adults and differed the most in the first year of age. However, higher CYP3A4 activity after birth was described by the Upreti model, and the expression reached similar levels as adults at 12 years old (Figure S1 ). This is associated with significantly lower predicted drug or metabolite exposures, particularly in those aged below 2 years old. The appropriate choice of a CYP3A4 ontogeny function remains controversial, and it is challenging to fully characterize due to the current limited pediatric data, especially in neonates and infants. Therefore, all tested maturation functions were accounted for while evaluating once‐daily dosing options in children. High exposures of DM‐6705 under the simulation setting of constant bioavailability using the ontogeny from Salem were predicted in those < 6 months of age at the start of the simulation, which may increase the risk of QT interval prolongation. However, the interim analysis from the PHOENIx trial evaluating delamanid PK in children concluded that both delamanid and DM‐6705 exposures were within the adult target range and a strong effect of age on DM‐6705 clearance is therefore unlikely. ²⁰ , ²⁴ Hence it is expected that the proposed once‐daily options for those aged below 6 months are acceptable, provided that patients are closely monitored for QTc prolongation throughout the treatment course.

There were some assumptions and limitations in our work. We assumed that the exposure‐QT relationship in children could be assessed using similar approaches as in adults, which is a standard way of QT evaluation. ⁴⁴ Distinct dietary intake patterns and time intervals between meal and drug administration could strongly influence the absorption of bedaquiline and delamanid. It has been reported that bedaquiline and delamanid exposures had 2‐fold and 2.7‐fold increases, respectively, while taken under fed state compared to fasting condition. ⁴ , ⁴⁵ In our simulation work, we assumed all patients were administered drugs together with food according to the labeled suggestions. Bedaquiline and delamanid are high protein‐binding drugs, and delamanid is thought to be predominantly metabolized by albumin. ³⁶ , ⁴⁶ Although albumin level has been found to increase with age, ⁴⁷ , ⁴⁸ Otsuka 232/233 pediatric TB trials showed no significant difference in albumin levels across the age spectrum. ¹⁹ Moreover, hypoalbuminemia was not identified as a covariate on clearance in previously developed delamanid PK models. ¹⁹ , ³⁴ In light of these reasons, albumin changes were assumed to have no considerable impact on bedaquiline and delamanid exposures and therefore not implemented in our simulation settings. In the delamanid pediatric model, dose‐dependent decreases on bioavailability and child‐friendly dispersible formulation effects on bioavailability were estimated from the adult population. ¹⁷ , ¹⁹ Whether these perform the same way as in pediatrics remains unclear. These factors, together with the variation in food intake, might jointly influence bioavailability, making them challenging to evaluate separately in children. A more comprehensive and mechanistic way to describe age‐specific changes in children is warranted. Given the assumptions made while implementing a model‐based exposure‐matching strategy for pediatric dose selection, data from ongoing and upcoming trials will be invaluable to re‐estimate these models and examine the proposed once‐daily dosing schemes under trial conditions.

This study utilizes model‐informed approaches to propose novel once‐daily dosing strategies for bedaquiline and delamanid in adults, adolescents, and children. Adjusted body weights with a TB disease correction factor, age‐dependent maturation process of the involved enzyme, along with child growth, were incorporated during simulation. Simpler and more feasible dosing regimens will allow fixed‐dose drug combinations and enhance patient adherence with the ultimate goal of TB elimination. The proposed once‐daily dosing regimens of bedaquiline and delamanid for adults are currently being tested in the PARADIGM4TB trial. In the pediatric version, age cutoffs at 3 and 6 months followed by weight‐based band dosing in those aged > 6 months were implemented. The dosing strategy in children will be evaluated in the IMPAACT 2020 study for all investigational drugs.

FUNDING

The project was funded from the National Institute of Health through the International Maternal Pediatric Adolescent AIDS Clinical Trials Network under award number UM1AI068632–17 for IMPAACT 2020 and the Innovative Medicines Initiative 2 Joint Undertaking (JU) under grant agreement No 101007873 for UNITE4TB. Overall support for the International Maternal Pediatric Adolescent AIDS Clinical Trials Network (IMPAACT) was provided by the National Institute of Allergy and Infectious Diseases (NIAID) with co‐funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) and the National Institute of Mental Health (NIMH), all components of the National Institutes of Health (NIH), under Award Numbers UM1AI068632 (IMPAACT LOC), UM1AI068616 (IMPAACT SDMC), and UM1AI106716 (IMPAACT LC), and by NICHD contract number HHSN275201800001I. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The JU receives support from the European Union's Horizon 2020 research and innovation programme and EFPIA, Deutsches Zentrum für Infektionsforschung e. V. (DZIF), and Ludwig‐Maximilians‐Universität München (LMU). EFPIA/AP contribute to 50% of funding, whereas the contribution of DZIF and the LMU University Hospital Munich has been granted by the German Federal Ministry of Education and Research. AC. Hesseling is also supported by a South African National Research Foundation SARChi Chair.

CONFLICTS OF INTEREST

E.M.S. and M.O.K. have received research grants from Janssen Pharmaceuticals. All other authors declared no competing interests for this work.

AUTHOR CONTRIBUTIONS

Y.‐J.L., L.E.L., M.O.K., A.J.G.‐P., A.C.H., and E.M.S. wrote the manuscript; Y.‐J.L., L.E.L., M.O.K., A.J.G.‐P., A.C.H., and E.M.S. designed the research; Y.‐J.L. performed the research; Y.‐J.L., L.E.L., M.O.K., A.J.G.‐P., A.C.H., and E.M.S. analyzed the data.

Supporting information

Data S1