Pharmacokinetics of anti-tuberculosis drugs in Venezuelan children younger than 16 years of age: supportive evidence for the implementation of revised WHO dosing recommendations

Authors


Corresponding Author Lilly M. Verhagen, Laboratory of Pediatric Infectious Diseases, Nijmegen, The Netherlands. E-mail: lillyverhagen@hotmail.com

Abstract

Objectives  The World Health Organization (WHO) recently issued revised first-line antituberculosis (anti-TB) drug dose recommendations for children, with dose increases proposed for each drug. No pharmacokinetic data are available from South American children. We examined the need for implementation of these revised guidelines in Venezuela.

Methods  Plasma isoniazid, rifampicin, pyrazinamide and ethambutol concentrations were assessed prior to and at 2, 4 and 8 h after intake of TB drugs by 30 TB patients aged 1–15 years. The effects of dose in mg/kg, age, sex, body weight, malnutrition and acetylator phenotype on maximum plasma drug concentrations (Cmax) and exposure (AUC0-24) were determined.

Results  25 patients (83%) had an isoniazid Cmax below 3 mg/l and 23 patients (77%) had a rifampicin Cmax below 8 mg/l. One patient (3%) had a pyrazinamide Cmax below 20 mg/l. The low number of patients on ethambutol (n = 5) precluded firm conclusions. Cmax and AUC0-24 of all four drugs were significantly and positively correlated with age and body weight. Patients aged 1–4 years had significantly lower Cmax and AUC0-24 values for isoniazid and rifampicin and a trend to lower values for pyrazinamide compared to those aged 5–15 years. The geometric mean AUC0-24 for isoniazid was much lower in fast acetylators than in slow acetylators (5.2 vs. 12.0, P < 0.01).

Conclusion  We provide supportive evidence for the implementation of the revised WHO pediatric TB drug dose recommendations in Venezuela. Follow-up studies are needed to describe the corresponding plasma levels that are achieved by the recommended increased doses of TB drugs.

Abstract

Objectifs:  L’Organisation Mondiale de la Santé (OMS) a récemment publié des recommandations posologiques révisées pour les antituberculeux (anti-TB) de première ligne pour les enfants, avec l’augmentation des doses proposées pour chaque médicament. Des données pharmacocinétiques ne sont pas disponibles pour les enfants sud américains. Nous avons examiné la nécessité de l’implémentation au Venezuela de ces recommandations révisées.

Méthodes:  Les concentrations plasmatiques d’isoniazide, de rifampicine, de pyrazinamide et d’éthambutol ont étéévaluées avant et à 2, 4 et 8 h après la prise des médicaments anti-TB par 30 patients TB âgés de 1 à 15 ans. Les effets de la dose en mg/kg, l’âge, le sexe, le poids, la malnutrition et le phénotype acétylateur aux concentrations plasmatiques maximales (Cmax) et l’exposition (AUC0-24) ont été déterminés.

Résultats:  25 patients (83%) avaient une Cmax d’isoniazide <3 mg/L et 23 patients (77%) avaient une Cmax de rifampicine <8 mg/L. Un patient (3%) avait une Cmax de pyrazinamide <20 mg/L. Le faible nombre de patients sous l’éthambutol (n = 5) a empêché des conclusions définitives. La Cmax et l’AUC0-24 de tous les quatre médicaments corrélaient significativement et positivement avec l’âge et le poids corporel. Les patients âgés de 1 à 4 ans avaient des valeurs Cmax et AUC0-24 significativement plus basses pour l’isoniazide et la rifampicine et une tendance à la baisse de ces valeurs pour la pyrazinamide par rapport à ceux âgés de 5 à 15 ans. La moyenne géométrique de l’AUC0-24 pour l’isoniazide était beaucoup plus faible chez les acétyleurs rapides que chez les acétyleurs lents (5.2 vs 12,0; p < 0.01).

Conclusion:  Nous fournissons ici des preuves d’appui pour l’implémentation des recommandations de l’OMS révisés pour les doses pédiatriques d’anti-TB au Venezuela. Des études de suivi sont nécessaires pour décrire les taux plasmatiques correspondants atteints avec les doses augmentées recommandées d’anti-TB.

Abstract

Objetivos:  La Organización Mundial de la Salud (OMS) ha presentado recientemente unas recomendaciones sobre las dosis de medicación antituberculosa de primera línea en niños, con un incremento de dosis propuesta para cada medicamento. No hay datos disponibles de farmacocinética en niños Sudamericanos. Hemos examinado la necesidad de implementar estas guías revisadas en Venezuela.

Métodos:  Se evaluaron los niveles en plasma de la isoniazida, rifampicina, pirazinamida y etambutol, midiendo, en 30 pacientes con edades entre los 1-15 años, las concentraciones antes de y después de 2, 4 y 8 h de haber ingerido los medicamentos anti-TB drugs. Se determinaron los efectos de dosis en mg/kg, la edad, sexo, peso, presencia de desnutrición y fenotipo acetilador en concentraciones máximas del medicamento en plasma (Cmax) y exposición (AUC0-24).

Resultados:  25 pacientes (83%) tenían Cmax para isoniazida por debajo de 3 mg/L y 23 pacientes (77%) tenían un Cmax para rifampicina por debajo de 8 mg/L. Un paciente (3%) tenía un Cmax para pirazinamida por debajo de 20 mg/L. El bajo número de pacientes con etambutol (n=5) restó firmeza a las conclusiones. La Cmax y AUC0-24 de los cuatro medicamentos estaban significativamente y positivamente correlacionadas con la edad y el peso. Los pacientes con edades entre 1-4 años tenían una Cmax y valores de AUC0-24 para isoniazida y rifampicina significativamente menores y una tendencia a disminuir los valores de pirazinamida al compararlos con aquellos con edades entre los 5-15 años. La media geométrica de AUC0-24 para isoniazida era mucho menor en acetiladores rápidos que en acetiladores lentos (5.2 vs. 12.0, P < 0.01).

Conclusión:  Ofrecemos evidencia que apoya la implementación de las recomendaciones revisadas de la OMS para la dosificación pediátrica de medicación para la TB en Venezuela. Se requieren estudios de seguimiento para describir los niveles en plasma alcanzados por las dosis incrementadas recomendadas.

Introduction

Tuberculosis (TB) in children is an important public health problem. Worldwide, at least half a million children become ill with TB and as many as 70 000 children die of TB each year (WHO 2012). The basic principles of treatment and recommended standard anti-TB regimens for children are similar to those for adults (WHO 2003). Until recently, dosage recommendations in mg/kg for these TB drugs in children were deduced from dosages used for adults without taking into account different distribution and clearance of these drugs in children. This may result in different plasma drug concentrations, which are supposed to be the link between the dose administered and response to TB drugs. Especially the total exposure to TB drugs (the area under the plasma concentration vs. time curve AUC0-24) and/or the peak plasma concentration (Cmax) appear to be relevant for TB drugs (Hall et al. 2009). WHO (2009) has recently issued revised dosage recommendations of the main first-line TB drugs for use in children. The recommended dosages and range of isoniazid, rifampicin, pyrazinamide and ethambutol are higher than previously recommended ones (Table 1). These revised recommendations are under consideration or being implemented in many settings across the world.

Table 1. Recommended first-line drug dosages for children as currently recommended and as previously recommended by WHO
DrugCurrently recommended (WHO 2009)Previously recommended (WHO 2003)
Daily dosage (dose range) in mg/kgDaily dosage (dose range) in mg/kg
  1. WHO, World Health Organization.

Isoniazid10 (10–15)5 (4–6)
Rifampicin15 (10–20)10 (8–12)
Pyrazinamide35 (30–40)25 (20–30)
Ethambutol20 (15–25)15 (15–20)

With all pharmacokinetic (PK) studies coming from Africa, Asia, Europe or North America (Zhu et al. 2002; Schaaf et al. 2005, 2009; McIlleron et al. 2009; Thee et al. 2009, 2010; Ramachandran et al. 2010), there is a lack of PK data of TB drugs in children from the South American region. As the pharmacokinetics of TB drugs, particularly isoniazid (Gumbo et al. 2007), may differ between different ethnic groups, the need for the revised dose recommendations and their effects on drug exposure should be examined in pediatric populations from different geographical regions. Furthermore, as the supply of TB drugs in middle income countries such as most South American countries is often complicated by difficulties in central purchasing and management (Chaulet 1992), it is important to justify the need for revised dose recommendations in order to create awareness to facilitate their implementation.

In Venezuela, the average national incidence of TB is moderate, between 25 and 50 per 100 000 (PAHO 2006). The present Venezuelan population is, like most South American populations, the product of crosses between Amerindians, Europeans and Africans (Martínez et al. 2007). We studied the pharmacokinetics of isoniazid, rifampicin, pyrazinamide and ethambutol in children at the Children’s Hospital ‘J.M. de los Ríos’ in Caracas Venezuela.

Methods

Patients

Between May and October 2011, 30 consecutive Venezuelan children aged 1–15 years on anti-TB treatment were investigated at the Department of Pediatric Infectious Diseases in the Caracas Children’s Hospital ‘J.M. de los Ríos’. Patients were included during the intensive phase of anti-TB treatment from two weeks up to two months after treatment initiation, when steady state concentrations are expected for all TB drugs. The diagnosis of TB was made on the basis of clinical signs and symptoms; gastric aspirate, sputum or other culture result positive for Mycobacterium tuberculosis; chest radiograph; a household source case with culture-positive TB and induration of ≥10 mm after a tuberculin skin test (TST) with 2 U of tuberculin RT23. Patients were excluded if they were older than 15 years of age, pregnant or lactating, had a history of liver or kidney disease, were HIV-positive [as these conditions may affect the PK of TB drugs (Graham et al. 2006; Schaaf et al. 2009)] or if the medical condition of the patient did not allow participation in the study as judged by the treating physician. HIV status was assessed in all patients upon diagnosis of TB as a routine investigation. Parents or legal guardians gave written informed consent before enrollment of their children. The study was approved by the Ethical Committee of the Instituto de Biomedicina, Caracas.

Drug treatment

Antituberculosis intensive phase treatment regimens in Venezuela at the time of research followed the previous WHO guidelines (Table 1). Daily isoniazid, rifampicin and pyrazinamide were administered as part of dispersible, pediatric, fixed-dose combinations (Kit Phase I: each tablet contains 60 mg of rifampicin, 30 mg of isoniazid, and 150 mg of pyrazinamide, manufactured by Cipla Limited, India).

Pharmacokinetic sampling and bioanalysis

Pharmacokinetic sampling took place in the Children’s Hospital J.M. de los Ríos in Caracas, Venezuela. Patients had taken their TB medication at 8 am during at least the 3 days before PK assessment. On the day of the PK assessment, TB drugs were taken right after a standardised breakfast (at 8 am) after an overnight fast from 11 pm the night before, which reflected the usual drug intake procedures in this population. The breakfast consisted of cachapas, typical Venezuealan low fat corn-based pancakes, with cheese and jelly and a glass of chocolate drink. Patients did not consume any other food for 4 h after TB drugs were taken. Serial venous blood samples were collected just prior to and at 2, 4 and 8 h after witnessed drug intake. Plasma was seperated by centrifugation within 30 min and was stored at −80 °C in Caracas and transported on dry ice to The Netherlands for bioanalysis. The concentrations of rifampicin, pyrazinamide and ethambutol were assessed with validated high-performance liquid chromatographic (HPLC) assays by methods described previously (Ruslami et al. 2007). Isoniazid and acetylisoniazid were measured with a validated method consisting of liquid-liquid extraction, followed by Ultra Performance Liquid Chromatography (UPLC) with ultraviolet (UV) detection. Alanine aminotransferase (ALT), aminotransferase (AST) and creatinine were determined as measures of liver and kidney function, respectively, at the day of PK sampling.

Pharmacokinetic analysis

Pharmacokinetic parameters were generated using the standard two-stage approach. With this approach, individual PK parameters are estimated in the first stage. In the second stage, population characteristics of each parameter are derived by obtaining measures of central tendency and spread of all the subjects’ individual parameters. PK parameters were assessed with non-compartmental methods using WinNonLin version 5.3 (Pharsight Corp., Mountain View, California). The concentration at 24 h post dose (C24) was calculated using the formula C24 = C8*e-β*(24−8), in which β is the first order elimination rate constant. β was obtained by least squares linear regression analysis on logC vs. time, with the absolute slope of the regression line being β/2.303. Half-life (t1/2) was obtained by the equation: 0.693/β. As a result of the small β or large t1/2 of pyrazinamide, β or a plasma t1/2 for this drug could not be estimated reliably based on sampling at 2, 4 and 8 h post dose. As patients had taken their TB medication at 8 am the day before the PK assessment and the sample T = 0 was taken at 8 am as well, the plasma concentration at time point T = 0 was assumed to reflect the concentration at 24 h post dose at steady-state. Therefore, to calculate β and t1/2 for pyrazinamide, we added a virtual time point T = 24 with the same plasma concentrations as measured on T = 0. AUC0-24 was calculated using the linear-log trapezoidal rule. Cmax and the sampling time of Cmax (tmax) were derived directly from the PK curves. Apparent clearance (Cl/F) was calculated by (dose/AUC0-24). The volume of distribution (Vd/F) was calculated by the equation (CL/F/[β]).

Determination of acetylator status

The PK of isoniazid is strongly dependent on the genetic polymorphism in N-acetyltransferase, a phase II metabolic enzyme (Aarnoutse 2011). We determined the acetylator status phenotypically from the half-life of isoniazid, as described in a population that included children (Miscoria et al. 1988). Patients with an isoniazid t1/2 less than 1.8 h were considered fast acetylators and those with a t1/2 greater than 1.8 h were classified as slow acetylators (Miscoria et al. 1988).

Malnutrition

The anthropometric measurements were transformed into weight-for-age, height-for-age, and body mass index (BMI)-for-age Z scores based on WHO standard reference populations (WHO 2006; De Onis et al. 2007) using WHO anthro software. Children under 5 years of age with height-for-age or weight-for-age Z scores <−2 standard deviations (SD) were defined as malnourished. Children aged 5–15 years with height-for-age or BMI-for-age Z scores <−2 SD were defined as malnourished, as weight-for-age is inadequate for monitoring growth beyond childhood (De Onis et al. 2007).

Statistical analysis

PK parameters (AUC0-24, Cmax, Vd/F, CL/F and t1/2) were log transformed before statistical analyisis and presented as geometric means and range. Tmax values were not transformed and were presented as median plus range. Comparison of AUC0-24 and Cmax between subgroups was performed using independent sample t-tests on the log-transformed parameters. Correlations between numerical variables were calculated using Pearson correlation on the log-transformed parameters. Correlations between PK parameters and doses were calculated using the non-parametric Spearman’s rank correlation coefficient (r). The SPSS programme for Windows version 16.0 (SPSS Inc, Chicago, IL, USA) was used for all statistical analyses.

Results

Thirty TB patients were enrolled. Patient and treatment characteristics are summarised in Table 2. Two children had elevated levels of both aspartate aminotransferase (ASAT, 114 U/l and 947 U/l) and alanine transaminase (ALAT, resp. 200 U/l and 599 U/l) at the day of PK sampling.

Table 2. Characteristics of 30 TB patients
CharacteristicValue
  1. *AFB, acid-fast bacilli; ALAT, alanine transaminase, ASAT, aspartate aminotransferase; IQR, interquartile range; SD, standard deviation; TB, tuberculosis.

Demographic characteristics
 Male, n (%)19 (63)
 Age, median (IQR)3 (2–8)
 Aged 1–4 years, n (%)17 (57)
 Aged 5–15 years, n (%)13 (43)
 Household contact, n (%)15 (50)
 TST positive (≥10 mm), n (%)18/25 (72)
 Hospitalised, n (%)4 (13)
 Number of days in anti-TB treatment, mean (SD)36 (15)
Type of TB
 Mycobacterium tuberculosis culture-positive or AFB-positive, n (%)6 (20)
 Pulmonary TB21 (70)
 Pleural TB2 (7)
 TB adenitis4 (13)
 TB meningitis1 (3)
 Spinal TB1 (3)
 Cutaneous TB1 (3)
Nutritional status
 Weight in kg, mean (SD)21 (13)
 Weight-for-age Z score in children <5 years, mean (SD)−0.65 (1.05)
 BMI-for-age Z score in children ≥5 years, mean (SD)0.15 (0.66)
 Malnourished, n (%)4 (13)
Laboratory values at day of pharmacokinetic analysis
 Creatinine, mean (SD)34 (12)
 ASAT in U/l, median (IQR)37 (29–48)
 ALAT in U/l, median (IQR)28 (23–34)
Dose in mg/kg, median (IQR)
 Isoniazid5.2 (4.6–5.6)
 Rifampicin10.4 (9.3–11.2)
 Pyrazinamide25.9 (23.2–28.0)
 Ethambutol17.2 (14.0–24.9)

Isoniazid

Twenty-five patients (83%) had peak plasma isoniazid levels below the reference range (3–6 mg/l, Table 3). None of the patients had a maximum concentration above the upper limit of the reference range. No significant differences were found in exposure (AUC0-24) between male and female patients (7.8 vs. 10.8, P = 0.30) or children who were malnourished compared to non-malnourished children (5.9 vs. 9.4, P = 0.22), although it should be noted that the statistical power may have been too low to find significant differences in the latter comparisons. There was a significant correlation between age and exposure (Pearson r 0.46, P = 0.01) and between body weight and exposure (Pearson r 0.47, P = 0.01). Exposure was significantly lower in children aged 14 years compared to children aged 515 years (AUC0-24 6.8 vs. 12.4, P = 0.02), while there was no significant difference in the percentage of fast and slow acetylators between the two age groups (P = 1.00). Similar results were obtained for Cmax (data not shown).

Table 3. Summary of PK parameters of TB drugs
Drugs and PK parameters*Isoniazid, = 30Rifampicin, n = 30Pyrazinamide, n = 30Ethambutol, n = 5
  1. CL/F, apparent clearance; PK, pharmacokinetic; TB, tuberculosis; Vd/F, volume of distribution.

  2. *Apart from tmax, for which median and range was displayed, geometric means and range were shown for all pharmacokinetic parameters.

  3. †Reference ranges in mg/l are 3–6, 8–24, 20–50 and 2–6 mg/l for isoniazid, rifampicin, pyrazinamide and ethambutol, respectively (Peloquin 2002). These reference values represent the normal concentrations that can be expected in adults after the standard doses of TB drugs. They are based on data that were compiled from all available sources (both healthy volunteers and TB patients) by, amongst others, Holdiness (Holdiness 1984) and Peloquin (Peloquin 1991). Subsequently, the ranges were validated in a range of phase I studies in healthy volunteers (Peloquin 2011).

AUC0-24 (h*mg/l)8.8 (2.6–31.0)20.6 (7.4–94.2)318.5 (108.0–1389.5)12.0 (5.4–28.2)
Cmax (mg/l)1.9 (0.7–5.2)5.1 (1.5–13.6)31.5 (10.3–71.7)1.8 (0.9–3.7)
% within reference range†17239460
tmax (h)2.0 (2.0–4.0)3.0 (2.0–8.0)2.0 (2.0–8.0)4.0 (2.0–4.0)
CL/F (l/h)10.3 (2.7–45.6)8.1 (2.5–17.1)1.4 (0.7–6.0)45.5 (36.1–61.2)
Vd/F (l)31.6 (13.9–85.3)16.0 (5.8–57.4)14.1 (5.5–50.2)216.2 (137.2–452.3)
t1/2 (h)2.1 (1.3–5.7)1.4 (0.4–4.6)6.8 (2.3–24.2)3.3 (2.5–7.4)

Neither AUC0-24 nor Cmax were significantly correlated with the dose of isoniazid in mg/kg (Spearman’s r 0.21, P = 0.27 and Spearman’s r 0.36, P = 0.05, respectively). The AUC0-24 and Cmax were strongly positively correlated (Pearson’s r 0.91, P < 0.01).

Acetylator phenotype

Eleven patients (37%) were considered fast metabolisers (t1/2 < 1.8 h) and 19 (63%) were slow metabolisers (t1/2 > 1.8 h). The geometric mean AUC0-24 was much lower in fast metabolisers than in slow metabolisers (5.2 vs. 12.0, P < 0.01, Figure 1). The Cmax was not significantly different between fast and slow metabolisers (1.5 vs. 2.1), P = 0.09). The two patients who had elevated ASAT and ALAT values were both slow acetylators. The Cmax of isoniazid was below the reference range in both patients and their AUC0-24 was not significantly different from the other patients.

Figure 1.

 AUC0-24 in fast (t1/2 < 1.8 h) and slow (t1/2 > 1.8 h) metabolisers. The boxplot shows the median AUC0-24 and the interquartile range. The whiskers represent the highest and lowest values.

Rifampicin

Twenty-three patients (77%) had peak plasma rifampicin levels below the reference range (8–24 mg/l). None of the patients had a maximum concentration above the upper limit of the reference range. Similar to isoniazid, no significant differences were found in exposure (AUC0-24) between male and female patients (18.8 vs. 24.5, P = 0.30) or children who were malnourished compared to non-malnourished children (20.7 vs. 20.6, P = 0.99). There was a significant correlation between age and exposure (Pearson r 0.48, P = 0.01) and between body weight and exposure (Pearson r 0.53, P < 0.01). Exposure was significantly lower in children aged 1–4 years compared to children aged 5–15 years (AUC0-24 16.2 vs. 30.1, P = 0.01). Similar results were obtained for Cmax (data not shown).

Both AUC0-24 and Cmax were significantly correlated with the dose of rifampicin in mg/kg (Spearman’s r 0.60 and 0.63, P < 0.01 for both correlations) and there was a strong positive correlation between AUC0-24 and Cmax (Pearson’s r 0.91, P < 0.01).

Pyrazinamide

One patient (3%) had a peak plasma pyrazinamide level below the reference range (20–50 mg/l). Another patient (3%) had a maximum concentration above the upper limit of the reference range. Again, no significant differences were found in exposure between male and female patients (AUC0-24 297.8 vs. 357.8, P = 0.23). A trend towards lower exposure in malnourished compared to non-malnourished children was observed (228.2 vs. 335.3, P = 0.07). As with isoniazid and rifampicin, there was a significant correlation between age and exposure (correlation coefficient Pearson r 0.44, P = 0.02) and between body weight and exposure (Pearson r 0.54, P < 0.01). Exposure was lower in children aged 14 years than in children aged 515 years, but this difference was not statistically significant (286.0 vs. 366.7, P = 0.09). Similar results were obtained for Cmax (data not shown).

Cmax, but not AUC0-24, was significantly correlated with the dose of pyrazinamide in mg/kg (Spearman’s r 0.53, P < 0.01 and Spearman’s r 0.32, P = 0.09, respectively). There was a strong positive correlation between AUC0-24 and Cmax (Pearson’s r 0.92, P < 0.01).

Ethambutol

Ethambutol was administered to only 5 (17%) of the patients in our study, as this drug may be omitted in children with non-cavitary, smear-negative pulmonary TB who are HIV-negative, in children with fully drug-susceptible TB and in young children with primary TB (WHO 2003). Two of the patients on ethambutol (40%) had a peak plasma ethambutol level below the reference range (2–6 mg/l). None of these patients had a maximum concentration above the upper limit of the reference range. None of the five patients were malnourished. Age and exposure as well as body weight and exposure were strongly correlated (Pearson r 0.94, P = 0.02 for both correlations). Exposure appeared to be lower in children aged 14 years (n = 2) than in children aged 515 years (n = 3) (5.4 vs. 20.2). Similar results were obtained for Cmax (data not shown).

AUC0-24, but not Cmax, was significantly correlated with the dose of ethambutol in mg/kg (Spearman’s r 1.00, P < 0.01 and Spearman’s r 0.80, P = 0.10). As with the other drugs, there was a strong positive correlation between AUC0-24 and Cmax (Pearson’s r 0.94, P = 0.02).

Discussion

This is the first report of PK characteristics of isoniazid, rifampicin, pyrazinamide and ethambutol in children from the South American region. One of our major findings, supporting the implementation of the revised WHO guidelines on dosage of TB drugs in children in Venezuela, is that more than three quarters of the children had isoniazid and rifampicin Cmax plasma levels below the reference ranges. Isoniazid and rifampicin concentrations in children in studies performed in other parts of the world have also reported to be lower than those in adults following the same mg/kg dose (Schaaf et al. 2005; McIlleron et al. 2009; Schaaf et al. 2009; Thee et al. 2009, 2010). Concerns have been raised about the possibility of drug-induced hepatotoxicity when doses of isoniazid will be increased. Two of the children in our study (7%) showed increased ASAT and ALAT values. In an extensive recent review of the literature, abnormal liver function tests were found in 10% of 3855 children on anti-TB treatment (Donald 2011). Although drug-induced hepatotoxicity occurs in children, the incidence is lower than in adults and the reviewer’s conclusion was that the use of higher dosages of isoniazid, rifampicin and pyrazinamide is unlikely to result in a greater risk of drug-induced hepatotoxicity in children.

Adequate peak plasma pyrazinamide concentrations were achieved in 28 of 30 children. This is in concordance with other studies showing that the lower recommended limit of serum pyrazinamide concentrations was generally achieved following the previous WHO recommendations in children (Gupta et al. 2007; Thee et al. 2008, 2011). However, in a recent study among adults, poor treatment outcome of pulmonary TB was associated with serum pyrazinamide concentrations of <35 mg/l (Chideya et al. 2009), suggesting that therapeutic pyrazinamide concentrations should be above the lower limit of 20 mg/l. Twenty-one (70%) of the children in our study had pyrazinamide concentrations below 35 mg/l.

The low peak plasma levels may be partly explained by the fact that Venezuelan children take their TB drugs just after breakfast, whereas PK reference levels are derived from studies in which patients refrained from food until 4 h after drug intake (Peloquin 2002). The intake of a meal with a high fat or carbohydrate content just before drug intake significantly reduces peak plasma levels of isoniazid and rifampicin, but not of pyrazinamide and ethambutol (Lin et al. 2010).

Values for total exposure (AUC0-24) may be even more relevant to the efficacy of first-line TB drugs than Cmax values (Hall et al. 2009). As commonly seen, there was a significant correlation between Cmax and AUC0-24 values for all first-line TB drugs in the current study. AUC0-24 values in our study can best be compared with those recorded in adult Indonesian TB patients (Ruslami et al. 2007) or Tanzanian TB patients (Tostmann et al., submitted), as the pharmacokinetics in these latter studies were assessed with the same analytical methodology at the same laboratory. Geometric mean exposures to isoniazid were 8.8 and 11.0 h*mg/l in Venezuelan children and Tanzanian adults, respectively. AUC0-24 values for rifampicin were 20.6, 39.9 and 48.5 h*mg/l in Venezuelan children, Tanzanian patients and Indonesian patients; pyrazinamide AUC0-24 values were 318.5, 344 and 472.8 h*mg/l respectively; and exposures to ethambutol were 12.0, 20.2 and 14.4 h*mg/l. These comparisons show lower exposures in Venezuelan children than adults, especially for rifampicin.

Maximum plasma concentrations of all four drugs were significantly correlated with age and body weight. AUC0-24 followed the same pattern as Cmax; children under 5 years of age had a lower AUC0-24 of isoniazid and rifampicin than children aged 515 years. Age-dependent elimination of isoniazid has been demonstrated, with younger children eliminating isoniazid more rapidly than older children (McIlleron et al. 2009).This has been attributed to a relatively larger liver size in proportion to total body weight in young children (Schaaf et al. 2005). In previous studies of serum pyrazinamide concentrations across different age groups, pyrazinamide pharmacokinetics did not differ significantly with age (Thee et al. 2008, 2011). Although we did not find a significantly lower pyrazinamide concentration in younger vs. older children, we did find a significant positive correlation of pyrazinamide peak plasma concentration and exposure with age, indicating that AUC0-24 and Cmax of pyrazinamide, as of the other three drugs, increase with age in our study population. Implementation of the revised dosage recommendations thus seems to be especially important for the youngest children in order to achieve satisfactory plasma concentrations.

Apart from age and body weight, the acetylator status for isoniazid appeared to be a strong determinant of exposure to isoniazid, as expected (Aarnoutse 2011). The metabolism of isoniazid occurs through acetylation by N-acetyltransferase 2 (NAT2), resulting in acetyl-isoniazid. NAT2 is characterised by a trimodal distribution that is genetically determined, resulting in fast, intermediate and slow metabolizers (Parkin et al. 1997). Phenotyping generally enables a discrimination between fast acetylators (including both the genotypic fast and genotypic intermediate acetylators) and slow acetylators (Aarnoutse 2011). Different proportions of fast and slow acetylator phenotypes have been reported based on ethnic or geographic population origin. About 50–60% of subjects from European (Caucasian), African and Indian origin are slow acetylators, whereas a minority varying from 5–25% among populations from Asian origin (Chinese, Japenese, Eskimos) is slow acetylator (Aarnoutse 2011). To our knowledge, our study is the first to describe acetylator phenotypes in a South American population. Our findings suggest that around 60% of Venezuelan children are slow acetylators, which roughly corresponds to European, African and Indian populations. The true number of children with a slow acetylator phenotype might be lower, as the percentage of children with a fast acetylator phenotype increases with age due to maturation of isoniazid acetylation over the first 4 years of life (Pariente-Khayat et al. 1997) and 57% of the children in our study sample were younger than 4 years of age.

Conclusion

This is the first PK study from the South American region focusing on pediatric TB. Low isoniazid and rifampicin plasma levels are shown following the previous WHO dose recommendations for children 115 years of age. Pyrazinamide concentrations were within the reference range. The small number of patients on ethambutol precluded firm conclusions about this drug. For isoniazid, rifampicin and pyrazinamide, exposure was lower among children aged 14 years compared to those aged 515 years. Our data provide supportive evidence for the implementation of the revised WHO guidelines for first-line anti-TB therapy in children in the South American region. Follow-up studies are needed to describe the exposures that are achieved when children take the currently recommended increased doses of TB drugs.

Acknowledgements

The authors thank Renzo N. Incani and the staff of the Hospital de Niños ‘J.M. de los Ríos’, Caracas (Caracas, Venezuela), the Laboratorio de Tuberculosis, Instituto de Biomedicina (Caracas, Venezuela) and the Department of Pharmacy, Radboud University Medical Centre (Nijmegen, The Netherlands) for their clinical assistance and technical support.

Ancillary