Pharmacokinetics, Safety, and Dosing of Novel Pediatric Levofloxacin Dispersible Tablets in Children with Multidrug-Resistant Tuberculosis Exposure

TEXT

Levofloxacin is increasingly important for the treatment and prevention of multidrug-resistant tuberculosis (MDR-TB) in adults and children. Levofloxacin is well absorbed after oral administration and eliminated primarily unchanged in the urine (1). Its pharmacokinetics in children have been evaluated in multiple studies with somewhat variable results (2–5). A major barrier to its use in young children to date has been lack of a widely available child-friendly formulation. In most high-TB-burden settings, adult 250-mg levofloxacin tablets are administered to young children after splitting or crushing them and mixing with water or food. The effect of this formulation manipulation on levofloxacin exposure is unknown.

A 100-mg scored dispersible taste-masked tablet has been developed (Macleods Pharmaceuticals, Ltd., Mumbai, India) and received World Health Organization (WHO) prequalification, with an expected increase in routine use for TB. The TB-CHAMP trial (ISRCTN92634082) is a phase 3 cluster randomized placebo-controlled trial to assess the efficacy of levofloxacin versus placebo for prevention of TB in child household contacts of adult MDR-TB source cases and supported development of this novel 100-mg levofloxacin formulation. As part of the TB-CHAMP trial, an open-label lead-in study in Cape Town, South Africa, was undertaken to characterize the pharmacokinetics and short-term safety of the levofloxacin 100-mg dispersible tablet in children aged <5 years.

Children aged <5 years were eligible for this lead-in study if they had household contact with an adult MDR pulmonary TB index case diagnosed during the previous 6 months, and written informed consent was provided by the parent or legal guardian. Key exclusion criteria in children included prevalent TB disease at enrollment, past receipt of preventive therapy with isoniazid or a fluoroquinolone for ≥16 weeks, TB treatment in the previous 12 months, and known concurrent exposure to an isoniazid-susceptible source case.

Levofloxacin 100-mg dispersible tablets were prescribed according to weight bands (target, 15 to 20 mg/kg once daily; actual, 10 to 21 mg/kg once daily) (see Table S1 in the supplemental material). Doses were prepared by placing the tablets in a plastic dosing cup or syringe, adding 2.5 to 10 ml of clean water, and swirling the cup or shaking the syringe until the tablets were fully dispersed and then administered. An additional 2.5 to 10 ml of clean water was added to the dosing container, swirled or shaken to suspend any remaining medication particles, and administered. Pharmacokinetic sampling was performed 7 to 14 days after the start of levofloxacin. An observed levofloxacin dose was given after an overnight fast, and samples were taken just before and at 1, 2, 4, 6, and 8 h after the dose. A standard breakfast was offered 1 h postdose. Levofloxacin concentrations were quantified by high-performance liquid chromatography with tandem mass spectrometry using a validated method as previously described (5).

All adverse events that occurred during the time the child was receiving the study levofloxacin formulation were recorded, assessed for attribution to levofloxacin, and graded for severity according to standard Division of AIDS grading (6).

Nonlinear mixed-effects modeling implemented in NONMEM (version 7.4) (7) was used to analyze data from the current study pooled with our previously published levofloxacin data (5). The previous study (MDRPK1) included 109 children with MDR-TB receiving the routinely available standard adult 250-mg solid tablet formulation (Austell, Johannesburg, South Africa), given as crushed tablets mixed in water and administered orally or by nasogastric tube, if the child refused to swallow it, or it was swallowed whole. The study population, site, sampling schema, sampling procedures, and laboratory assays were similar to those of the current study. The pharmacokinetics model from the previous study was used as a starting point, but the model structure was revisited and the model parameters reestimated when fitting the new data. Improvements in objective function value (ΔOFV) and goodness-of-fit plots were used for model development. The final model was then used to optimize weight-band dosing targeting the levofloxacin area under the curve (AUC) achieved in adults with TB and normal renal function receiving a 750-mg daily dose, set as 96.8 mg·h/liter. This target was based on rescaling previously reported values in adults with TB receiving a 1,000-mg dose, because levofloxacin pharmacokinetics have been shown to be linear in this dose range (5, 8).

Informed consent was provided by parents or legal guardians. Ethics approval for the study was provided by the Health Research Ethics Committees of Stellenbosch University (M16/02/009).

Twenty-eight children were enrolled. One child was withdrawn before pharmacokinetic sampling, so 27 children contributed to safety data; however, 3 children were unable to complete pharmacokinetic sampling because of difficulties with phlebotomy. The median age of children completing pharmacokinetic sampling (n = 24) was 2.1 years (interquartile range, 1.2 to 2.7 years); none were infected with HIV, 3 (13%) had a weight-for-age Z-score of less than −2. Complete baseline characteristics are shown in Table 1, along with baseline characteristics of the historical MDRPK1 study population.

A total of 144 levofloxacin concentrations were available for analysis, with two predose samples below the limit of quantification. The levofloxacin pharmacokinetic model structure was the same as and the parameter values not significantly different from those of the previous analysis, as shown in Table 2, except for the bioavailability of the new dispersible tablets. When the new dispersible tablet was chosen as reference for biavailability (F = 1), the tablets used in the previous study were found to be 41% less bioavailable (ΔOFV = 132, 1 extra degree of freedom; P < 10⁻⁶). Visual predictive checks are provided in Fig. 1, which shows how the model suitably fit the data of both the previous and current studies and the large difference in exposure due to the change in bioavailability. The value of clearance (CL)/F for a typical 12-kg 2-year-old child was 2.8 liter/h. After adjusting for full maturation and using allometric scaling to resize to a 70-kg adult, the value became 11.4 liters/h, which was more consistent with scaled values from published adult (CL/F, 8.35 to 9.16 liters/h) and pediatric (CL/F, 11.6 to 15.4 liters/h) studies that we previously summarized compared to the standard 250-mg adult tablet we previously studied (CL/F, 18.8 liters/h) (5). Interestingly, administration by nasogastric tube was not found to affect the speed of absorption of the dispersible tablets, unlike results with the standard 250-mg tablet in the previous study, for which the absorption lag time was shortened when a nasogastric tube was used for drug administration.

• Open in new tab
• Download powerpoint

FIG 1

Visual predictive check of the levofloxacin concentration versus time after dose for the historical 109 controls receiving adult 250-mg standard levofloxacin tablets (5) (left) versus the 24 patients on the pediatric 100-mg scored dispersible tablets (right). Solid and dashed lines, 50th, 5th, and 95th percentiles of the observed data; shaded areas, model-predicted 95% confidence intervals for the same percentiles. Circles, observed concentrations.

Proposed weight-banded dosing with this novel dispersible tablet is shown in Table 3. Despite the higher exposures in children receiving the dispersible tablets compared with those in children receiving the adult formulation, doses of >15 to 20 mg/kg in some weight bands were still required to achieve adult target exposures. Simulated maximum plasma concentrations (C_max) and AUC from 0 to 24 h (AUC_0–24) with this dosing approach are shown in Fig. 2.

• Open in new tab
• Download powerpoint

FIG 2

Simulated steady-state levofloxacin C_max and AUC_0–24 versus body weight using weight-banded dosing of pediatric levofloxacin 100-mg scored dispersible tablets. The dashed line for AUC (96.8 mg · h/liter) is the median AUC from adults with TB receiving 1,000 mg daily in the study by Peloquin et al. (8) after linearly rescaling the dose from 1,000 to 750 mg/day, because the pharmacokinetics of levofloxacin have been reported to be linear in this dose range. The dashed line for C_max (15.55 mg/liter) is the median C_max in adults receiving 1,000 mg/day from the same study. C_max was not rescaled to a 750-mg/day dose, because the C_max values achieved with the 1,000-mg dose were safe in adults over 7 days in that study. In selecting doses for children, it is helpful to know that the C_max values expected in children with the proposed doses are not well beyond what has been demonstrated to be safe in adults.

Two participants had three adverse events at least possibly related to levofloxacin, including vomiting (grade 1), difficulty sleeping (grade 1), and anorexia (grade 2). All adverse events and those at least possibly related are shown in Tables S2 and S3 in the supplemental material. No patients had grade 3/4 adverse events or any serious advents, and none discontinued levofloxacin because of potentially drug-related adverse events.

Use of this novel pediatric levofloxacin dispersible tablet resulted in substantially higher exposures in children relative to the same doses of adult 250-mg solid tablets used in previous pediatric pharmacokinetic studies by our group (5), with bioavailability primarily accounting for this difference. The TB-CHAMP and MDRPK1 populations were different in a number of ways other than type of formulation received, including MDR-TB disease status (MDR-TB disease versus exposure), ethnicity, and use of a nasogastric tube for medication administration. Our previously published work (MDRPK1) included children receiving levofloxacin either for MDR-TB treatment, often in combination with at least five other TB medications, or for MDR-TB preventive therapy, given in combination with isoniazid and ethambutol. We cannot rule out an interaction with other TB drugs that might explain the difference in exposures between the two studies; however, this is unlikely, since no such interactions were previously described. This may be worth evaluating in future studies. Additionally, inclusion of disease status (MDR-TB disease versus exposure) did not improve the model fit, which is also reassuring. Ethnicity was evaluated and not found to improve the model fit, so it is unlikely to explain the observed differences in exposures. The effect of administration by nasogastric tube was evaluated in the original MDRPK1 analysis and found to increase the rate of absorption but not affect the overall bioavailability. This makes it less likely that the difference between the studies in nasogastric tube administration explains the differences in levofloxacin exposures. In the previous analysis of the MDRPK1 study and in the current pooled analysis, HIV infection was associated with a 16% lower CL, resulting in higher exposures. The effect of HIV was included in the final model, so the model would describe the difference in bioavailability after accounting for the HIV effect on CL. This is important to note even though there was no statistical difference in the proportion of HIV-infected children in the two studies. The difference in bioavailability that we describe is likely due to the effect of formulation. However, the differences between the populations receiving the two different formulations are a limitation of this analysis, and future work specifically designed to assess the effect of formulation would improve the confidence in our findings.

The effect of formulation is increasingly being recognized as an important contributor to drug exposure, and it may differ between children and adults (9, 10). The value for levofloxacin’s apparent oral clearance (CL/F) with the 100-mg dispersible tablets, after adjusting for body size with allometric scaling, is lower than we described previously with the 250-mg immediate-release tablets but consistent with most previously published values (5). Moreover, when choosing the bioavailability of the new formulation as reference (value = 1), and thus rescaling the bioavailability of the previous study to 0.587, the disposition parameters of the previous analysis become consistent with the historical reports on levofloxacin pharmacokinetics. This suggests that the routinely used levofloxacin formulation studied in our previous work (and used in routine TB care in South Africa) had unexpectedly low bioavailability. Additional work to understand potential formulation effects is warranted given the widespread global use of the 250-mg levofloxacin tablet in children with TB.

The lack of safety concerns in this study is reassuring and consistent with previous reports (11). However, careful safety monitoring should continue in children receiving levofloxacin, particularly with longer treatment durations; the main TB-CHAMP trial will contribute such data.

The levofloxacin formulation reported here is expected to be used widely by national TB programs for treatment of children in routine settings. Our proposed weight-banded doses can be adopted by programs for dosing in children with this formulation. International dosing recommendations for children with TB should be updated for levofloxacin, taking into consideration emerging data and new formulations for levofloxacin, as is needed for other second-line anti-TB medications.