Population/ Nonlinear mixed-effects modelling of pharmacokinetics and pharmacodynamics of tuberculosis treatment

Doctoral Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title

University of Cape Town

The pharmacokinetics of rifampicin, isoniazid, pyrazinamide and ethambutol in TB/HIV coinfected patients recruited in two phase III clinical trials (61 patients in TB-HAART and 222 patients in RAFA study) were described using nonlinear mixed-effects modelling. Concentration-time data for rifampicin (TB-HAART study) was used to develop a semimechanistic pharmacokinetic model incorporating autoinduction and saturable pharmacokinetics. A model describing the pharmacokinetics of pyrazinamide (TB-HAART study) was developed and used to evaluate the 24-hour area under the concentration-time curve (AUC0–24), and maximum concentrations (Cmax) achieved with the currently recommended weight-adjusted doses for drug-susceptible and -resistant tuberculosis. Concentration-time data from the RAFA study were used to characterise the pharmacokinetics of the four drugs of the fixed dose combination (FDC) therapy including desacetyl-rifampicin, and acetyl-isoniazid. Binary recursive techniques were applied in the conditional inference framework to determine predictors including drug exposure of time-to-stable culture conversion and poor long-term treatment outcomes. The model describing the pharmacokinetics of rifampicin predicted that increasing the dose results in a more than proportional increase in exposure. Clearance of rifampicin increased by 90% from baseline to steady-state due to autoinduction and the process takes up to 21 days. Monte Carlo simulations showed that rifampicin doses of at least 25 mg/kg would be required to achieve an AUC0–24/MIC ratio of at least 271. Based on the model describing the pharmacokinetics of isoniazid, co-administration of isoniazid and efavirenz-based antiretroviral therapy results in a 54% reduction in isoniazid exposure only in fast acetylators. There were disparities in exposure across weight bands for all the four drugs: patients with lower weight had reduced exposure. To match drug exposure across the weight bands, we recommend the addition of one FDC tablet to patients with weight less than 55 kg. There is need to explore the use of fat-free mass-adjusted dosing since cumulative evidence shows its superiority over total body weight in driving exposure via allometric scaling for all first-line antituberculosis drugs. Individual drug exposures were not predictive of either time-to-stable culture conversion or long-term tuberculosis treatment outcomes. Baseline X-ray grading, HIV stage as TB diagnosis, and treatment arm were predictive of time-to-stable culture conversion while the presence of cavities, patient’s level of physical activity and CD4 count were the drivers of long-term treatment outcomes.