Article Text

Download PDFPDF

Original article
Pharmacokinetic study of isoniazid and pyrazinamide in children: impact of age and nutritional status
  1. Rajeshwar Dayal1,
  2. Yatish Singh1,
  3. Dipti Agarwal2,
  4. Manoj Kumar1,
  5. Soumya Swaminathan3,
  6. Geetha Ramachandran4,
  7. Santosh Kumar5,
  8. Shamrendra Narayan6,
  9. Ankur Goyal7,
  10. A K Hemant Kumar4
  1. 1 Department of Pediatrtics, Sarojini Naidu Medical College, Agra, Uttar Pradesh, India
  2. 2 Department of Paediatrics, Dr. Ram Manohar Lohia Institute of Medical Sciences, Lucknow, Uttar Pradesh, India
  3. 3 Indian Council of Medical Research, New Delhi, India
  4. 4 National Institute for Research in Tuberculosis, Chennai, India
  5. 5 Department of TB & Chest Diseases, Sarojini Naidu Medical College, Agra, Uttar Pradesh, India
  6. 6 Department of Radiology, Dr. Ram Manohar Lohia Institute of Medical Sciences, Lucknow, Uttar Pradesh, India
  7. 7 Department of Microbiology, Sarojini Naidu Medical College, Agra, Uttar Pradesh, India
  1. Correspondence to Dr Dipti Agarwal, Faculty Residential Apartment, Dr Ram Manohar Lohia Institute of Medical Sciences, Lucknow 226010, India; drdiptiagarwal{at}


Objectives To evaluate pharmacokinetics of first-line antitubercular drugs, isoniazid (INH) and pyrazinamide (PZA), with revised WHO dosages and to assess its adequacy in relation to age and nutritional status.

Design Observational study.

Setting This study was conducted at Sarojini Naidu Medical College, Agra, and National Institute for Research in Tuberculosis, Chennai.

Patients 40 subjects diagnosed with tuberculosis were registered in the study and started on daily first-line antitubercular regimen based on the revised WHO guidelines.

Interventions Blood samples were collected at 0, 2, 4, 6 and 8 hours from these subjects after 15 days of treatment for drug estimations.

Main outcome measure The measurement of drug concentrations (maximum peak concentration (Cmax) and area under the time –concentration curve (AUC0–8 hours)) for INH and PZA. Appropriate statistical methods were used to evaluate the impact of age and nutritional status on pharmacokinetic variables.

Results For INH, the difference in drug exposures in children <3 years (Cmax 3.18 µg/mL and AUC0–8 hours15.76 µg/mL hour) and children >3 years (Cmax3.05 µg/mL and AUC0–8 hours 14.37 µg/mL hour) was not significant (P=0.94, P=0.81, respectively). The drug levels in children with low body mass index (BMI) (Cmax3.08 µg/mL; AUC0–8 hours14.81 µg/mL hour) were also comparable with their normal counterparts (Cmax3.09 µg/mL, P=0.99; AUC0–8 hours 14.69 µg/mL hour, P=0.82). PZA drug exposures obtained in children less than 3 years (Cmax29.22 µg/mL, AUC0–8 hours 155.45 µg/mL hour) were significantly lower compared with drug levels in children above 3 years (Cmax 37.12 µg/mL, P=0.03; AUC 202.63 µg/mL hour, P value=0.01). Children with low BMI had significantly lower drug concentrations (Cmax 31.90 µg/mL, AUC0–8 hours167.64 µg/mL hour) when compared with normal counterparts (Cmax 37.60 µg/mL, P=0.02; AUC0–8 hours 208.77 µg/mL hour, P=0.01).

Conclusions The revised WHO drug dosages were found to be adequate for INH with respect to age and nutritional status, whereas PZA showed significantly lower drug levels in children <3 years and in malnourished children.

  • tuberculosis
  • children
  • pharmacokinetics
  • antituberculosis drugs
  • malnutrition

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic?

  • As therapeutic concentrations of antitubercular drugs were not achieved in younger and malnourished children using the previously recommended dosages of isoniazid and pyrazinamide, WHO has upgraded the dosage sof these drugs in children (INH 10(10–15) mg/kg, PZA 35(30–40) mg/kg).

What this study adds?

  • PZA drug peak concentrations (maximum peak concentration (Cmax)) and drug exposure (area under the time – concentration curve (AUC0–8)) were inadequate in children under 3 years of age and in those with low body mass index using the revised WHO dosages.

  • Age and nutritional status did not have impact on drug peak concentrations (Cmax) and exposure (AUC0–8) with INH.


Tuberculosis (TB) remains a major global health problem. WHO reported that in 2015, 1 million children were suffering from TB globally (<15 years), with a mortality of 136 000 each year.1 Various studies on pharmacokinetics of antitubercular drugs in children have shown subtherapeutic drug levels on using the absolute adult doses in children.2–7

Children absorb, metabolise and eliminate these drugs in a different manner than adults. It is asimple extrapolation of adult doses to children, without correctly accounting for the changes in pharmacokinetics with age and nutritional status that may lead to incorrect dosing in children. On the basis of these findings, WHO guidelines for treatment were revised in 2010 for isoniazid (INH 10, 7–15 mg/kg), rifampicin (RMP 15, 10–20 mg/kg), pyrazinamide (PZA 35, 30–40 mg/kg) and ethambutol (E 20, 15–25 mg/kg).

Variations in body weight, composition, surface area along with enzyme maturation in children need to be taken into consideration while calculating doses of antitubercular drugs.8 These changes are maximally seen in children up to 2 years of age. Young children are known to eliminate INH at a faster rate due to the larger mass of the liver in proportion to total body weight.9 Drugs used previously were calculated on the basis of age or body weight, which would be inadequate for children with poor nutritional status.10 Drug metabolism and malnutrition are closely related to each other.11 Poor nutritional status may influence the pharmacokinetics by causing decreased absorption, gastrointestinal infections, decreased levels of enzymes, changes in the protein binding of drugs and altered renal functions.12 13

Hence, this pharmacokinetic study was conducted to study the drug levels of INH and PZA in children with revised WHO dosages in children and to assess the effect of nutritional status and age on the pharmacokinetics variables of these drugs. We studied the pharmacokinetics of INH and PZA since WHO has revised the doses of INH and not many studies have been conducted with the revised doses of INH and PZA.4 5


Children aged 1–15 years attending Outpatient Department or admitted in Paediatrics Department and TB Chest Department, Sarojini Naidu Medical College, diagnosed with TB were included in the study. Diagnosis was made in accordance with the National Consensus Guidelines, 2012.14 Children whose parent/guardian gave consent for the study were included in the study. Assent was taken from children older than 7 years. The study included all forms of TB in paediatric subjects. Children with HIV infection, sick/moribund patients and patients with clinical evidence or a history of renal or liver disease or any other condition requiring a modification of anti-TB treatment were excluded from the study. A thorough clinical examination was done. Classification of malnutrition was done using WHO growth standards for children. Laboratory investigations included complete blood count, liver function tests and renal function tests, chest X-ray and tuberculin skin test. Acid fast bacilli microscopy and solid culture from relevant clinical specimen (sputum, cerebrospinal fluid, ascitic fluid and pleural fluid) were performed in all subjects. Ultrasonography  abdomen, CT scan/MRI scan was done as required. The pharmacokinetic study was conducted only after anti-TB treatment had been taken regularly for at least 15 days. Eligible patients were admitted to the ward a day prior to the study day. On the day of the study, a sample of blood (2 mL) was collected under fasting condition in a heparinised Vacutainer tube, followed by administration of anti-TB medications. The drugs were administered at the revised WHO doses under supervision. Blood samples (2 mL) were collected at 2, 4, 6 and 8 hours in heparinised Vacutainer tubes after drug administration. The children had regular breakfast and lunch under supervision. The children were discharged from the hospital after the 8-hour blood sample was drawn. The children were followed up upto 6 months of antitubercular therapy. The blood samples were centrifuged immediately after collection and plasma was stored at –20°C until estimation of anti-TB drugs was undertaken. Plasma concentrations of INH and PZA were measured by high- performace liquid chromatography.15 Calibration curves with concentrations ranging from 0.1 to 10.0 µg/mL for INH and 5.0– 50.0 µg/mL for PZA were prepared in drug-free human plasma. 6-Aminonicotinic acid was used as an internal standard. Unknown drug concentrations were derived from linear regression analysis of the peak height ratios (analyte/internal standard) versus concentration curve.

Statistical analysis

Based on the plasma concentrations of INH and PZA obtained at different time points, certain pharmacokinetic variables were calculated. The maximum peak concentration (Cmax) was determined by visual inspection of data. Exposure or area under the time–concentration curve (AUC0–8) was calculated by non-compartmental analysis. The categorical variables were assessed using Χ2 test and continuous variables were analysed using Wilcoxon rank-sum test. The age of study subjects was stratified into two groups: <3 years and >3 years. Subtherapeutic Cmax values (INH <3 µg/mL, PZA <35 µg/mL) were defined as per the National Jewish Medical Research Centre study.16 Univariate and multivariable regression analyses were performed to evaluate factors (age and nutritional status) that influenced pharmacokinetic variables. We calculated the power of the study with respect to the differences obtained by age and body mass index (BMI) and found to be more than 80%.


In the present study, 40 subjects were registered and started on first-line antitubercular therapy based on revised WHO drug dosages. On drug estimation, samples of three subjects were found to be indeterminate, thus results of 37 subjects were analysed. The profile of study subjects is given in table 1. The study subjects were predominantly male 27 (73%) with median age of 8 years (IQR 3–10 years). According to WHO standards for weight for height (WHZ), 11 (52%) subjects were severely wasted. Weight for age (WAZ) showed that 19 (66%) subjects were underweight. Based on height for age (HAZ) scores, 15 (37.5%) subjects were found to be stunted. Based on BMI, 14 (35%) subjects were severely malnourished as shown in table 1. Most of the study participants enrolled in the study had extrapulmonary TB (51.4%).

Table 1

Clinical profile of study subjects (n=37)

Pharmacokinetic details of INH

The mean Cmax and AUC0–8 hours for INH were 3.08±1.72 µg/mL and 14.74±9.64 µg/mL hour, respectively. The Cmax in children <3 years was 3.18 µg/mL and for children >3 years was 3.05 µg/mL. There was no statistically significant difference (table 2A) between the two values observed. The AUC0–8 hours for children <3 years was 15.76 µg/mL hour and for children >3 years was 14.37 µg/mL hour. There was no statistically significant difference between the two observed values (table 2B). On considering nutritional status, as mentioned in table 3, the Cmax for underweight (WAZ < −2), stunted (HAZ < −2), wasted (WHZ <  −2) and malnourished (BMI < −2SD) were not statistically different when compared with their normal counterparts. Similarly, AUC0–8 hours for underweight, stunted, wasted and malnourished were not statistically different when compared with their normal counterparts. In multivariate regression analysis, the type of TB influenced the peak concentration and exposure of INH. Children with extrapulmonary TB exhibited lower Cmax and AUC0–8 hours compared with pulmonary TB (P=0.06, P=0.05), respectively.

Table 2

Factors influencing

Table 3

Correlation of nutritional status and pharmacokinetic variables

Pharmacokinetic details of PZA

The mean Cmax and AUC0−8 hours for PZA were 34.98±9.75 µg/mL and 189.87±57.21 µg/mL hour, respectively. The Cmax for children <3 years was 29.22 µg/mL and for >3 years was 37.12 µg/mL, which was statistically significant (P=0.031). The AUC0−8 hours for children <3 years was 155.45 µg/mL hour, which was statistically lower than observed in children and for >3 years (202.63 µg/mL hour) as shown in table 2B. The Cmax for children with low BMI (31.90 µg/mL) was significantly lower (table 3) than for children with normal BMI (37.60 µg/mL). No statistically significant difference in Cmax was observed between normal children and children with WAZ, WHZ and HAZ scores below −2SD. However, the AUC0−8 hours for children with HAZ scores and WHZ scores below −2SD was significantly lower than children with normal scores. The AUC0−8 hours for children with low BMI was 167.64 µg/mL hour while for children with normal BMI was 208.77 µg/mL hour showing statistically significant difference. Children with extrapulmonary TB had Cmax values of 36.76 µg/mL and children with pulmonary TB had Cmax values 36.90 µg/mL. No statistically significant difference was observed (P=0.63). The AUC0−8 hours value for children with extrapulmonary TB was 195.85 µg/mL hour and the AUC 0−8 hours value for children with pulmonary TB was 206.66 µg/mL hour, which was not statistically significant (P=0.50).

The study was adequately powered to detect differences in drug concentration among different groups of children. Out of the 37 patients who were assessed after 6 months of therapy, 35 (94.6%) had a favourable outcome, while 2 (5.4%) were lost to follow-up.


The study was aimed to find the impact of age and nutritional status on the pharmacokinetic variables in children given antitubercular drugs at revised WHO dosage regimens. The two main variables studied were Cmax and AUC0−8 hours for INH and PZA.


In our study, the Cmax and AUC0−8 hours levels for INH were not influenced by age and nutritional status. Similar results were observed by Mukherjee, Thee and Kwara et al (using revised WHO doses).17–19 Geetha et al, using the older dose regimes, found that the Cmax and AUC0–8 of INH were significantly lower in children aged 1–3 years than in the other age groups (P<0.01).20 The study by Schaaf et al in children with TB showed that INH was eliminated faster in younger children compared with older children, thus achieving lower concentrations.9 Younger children are reported to have lower drug levels, despite receiving similar drug doses on mg/kg basis as older children. This is probably due to faster elimination of drugs by infants and younger children because of a relatively greater mass of the liver in proportion to total body weight.

Cmax and AUC0−8 hours values of INH in our study reflected no variation with respect to malnutrition, which again corroborates with findings of Mukherjee, Thee and Kwara et al.17–19 Similar findings have been reported by Seifart et al who showed that alteration in nutritional status in children did not have effect on INH metabolism in children.21 Eriksson et al have shown that INH achieved adequate drug levels in children with TB, who were in different nutritional categories (normal, underweight, marasmus and kwashiorkor).22 On the other hand, Geetha et al observed that the Cmax and AUC0–8 concentrations of INH were significantly lower in stunted children.20 Some earlier studies by Seth et al on INH have shown that plasma concentrations of these drugs were more in children with poor nutritional status.23 Drug metabolism and nutritional status have a close association with each other. The metabolic changes that occur in relation to nutritional status can change pharmacokinetic processes, drug responses and their side effects.11 In our study, subjects with extrapulmonary TB had significantly lower Cmax values when compared with pulmonary TB cases. No such correlation has been reported in previous pharmacokinetic studies.


In our study, both age and nutritional status had a significant influence on the drug concentrations even with revised drug dosages. A significantly higher Cmax and AUC0-8hr values in children >3 years were observed. Similar observations were made by Geetha et al, who found that Cmax and AUC0–8 hours of PZA were significantly lower in children aged 1–3 years than in the other age groups.20 Kwara et al also found significantly lower Cmax and AUC0–8 hours values in children <2 years when compared with >2 years.19 This is probably due to faster elimination of drugs by infants and younger children due to the relatively greater mass of the liver in proportion to total body weight. Differences in absorption, distribution, metabolism and excretion due to growth and development in children are known to influence plasma drug concentrations.24 On the contrary, studies of Mukherjee and Thee et al (using revised WHO dosage recommendations) found no such relationship with age.17 18 In our study, malnourished (low BMI) children had significantly lower Cmax and AUC0–8 hours values. The AUC0–8 hours values in our study were significantly lower in stunted and wasted children. Geetha and Graham et al also observed a significant lower Cmax values and AUC0–8 hours values in stunted and underweight children.20 Kwara et al found that stunted children in their study had lower median Cmax and AUC0–8 hours values.19 In a systemic review, Oshikoya et al showed that malnutrition in children resulted in lower plasma levels of antitubercular drugs, especially PZA. Malnutrition may influence pharmacokinetics by causing malabsorption, enteric infections, altered levels of enzymes, changes in the protein binding of drugs, altered renal functions and changes in the body composition.12 13 Mukherjee et al and Thee et al found no significant effect of malnutrition on pharmacokinetic variables of PZA.17 18

The study did not observe any deaths as sick/moribund patients were excluded from the study. Most of the subjects had favourable outcome in our study despite subtherapeutic drug levels of PZA in younger and malnourished children. It could be probably due to synergistic action of PZA with other antitubercular drugs, which may reduce the effect of lower levels of one drug.25 It should be emphasised that the relationship of treatment outcome with drug levels remains a complex one. In addition to plasma drug levels, treatment outcome depends on various factors, namely bacillary load, type and virulence of strain, minimal inhibitory concentration, drug concentrations at the site of lesion and duration of infection. Treatment outcome is also guided by the severity of disease, immune status and nutritional status of the subject.

A major study limitation was that we could not include RMP concentrations in this study due to small volumes of plasma that prevented from estimating this drug. The small numbers of subjects in the study deterred the in-depth assessment of association of serum drug concentrations with age and nutritional status.


Higher doses of INH in revised WHO protocols have shown adequate drug levels, especially in younger children and in children with malnutrition. However, increasing the dose of PZA did not significantly increase Cmax and AUC0–8 hours concentrations achieved in children <3 years of age. Poor nutritional status showed close relationship with low drug levels of PZA. The results of this study have raised further questions regarding adequacy of the revised PZA dosages and thus add to the dilemma whether achieving therapeutic plasma concentrations or clinical improvement be made the goal of adequate treatment. While changing the doses of anti-TB drugs in children, other factors which may be influencing plasma levels and therapeutic outcomes should be considered, keeping in mind the dreaded drug resistance with inappropriate treatment.



  • Contributors RD and SS conceived the idea of the study. YS, DA, SN, SK and AG were involved in data collection. DA, GR and MK were involved in final interpretation of the result. GR and AKHK supervised the pharmacokinetic study of the samples and edited the manuscript. DA, YS and SN wrote the first draft of the manuscript. The final manuscript was approved by all authors.

  • Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

  • Patient consent Parental/guardian consent obtained.

  • Ethics approval Institute’s ethical committee, Sarojini Naidu Medical College, Agra India.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles