Effect of rifampicin-based antitubercular therapy and cotrimoxazole on the population pharmacokinetics of stavudine (d4T) in HIV-1 infected patients
J-S van der Walt , K Cohen , HM McIlleron , PJ Smith , G Maartens , MO Karlsson 
 Division of Clinical Pharmacology, University of Cape Town, Cape Town, South Africa,  Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden
Objectives: The coadministration of antiretroviral and antitubercular therapy is known to result in drug-drug interactions due primarily to the induction of metabolising enzymes and drug transporters by rifampicin. Nucleoside analogue reverse transcriptase inhibitors, including stavudine (d4T), are predominantly eliminated by renal tubular secretion and not affected by this interaction. The aims of our analysis were to describe the population pharmacokinetics of stavudine during and after antitubercular treatment, to assess the effects of cotrimoxazole on stavudine pharmacokinetics and to quantify the interoccasional variability in this population.
Methods: Stavudine concentration-time data from 16 patients who received stavudine-containing antiretroviral therapy during and after rifampicin-based antituberculosis therapy were analysed using nonlinear mixed effects modelling (NONMEM version VI, FOCE). The effects of antitubercular therapy were assessed on both fixed effects (the relative bioavailability, i.e. during vs after antitubercular therapy, and the first order absorption rate constant) and random effects (interoccasional variability).
Results: A one-compartment disposition model with first order absorption and elimination, and absorption lag time best described the data. The pharmacokinetics of stavudine was highly variable and the base model included between-subject variability (BSV) on the apparent clearance (CL/F), absorption rate constant (Ka) and relative bioavailability. Antitubercular therapy did not have a significant effect the relative bioavailability or absorption of stavudine. During antitubercular treatment the bioavailability was 3% lower (proportional change: -0.03; 95% CI: -0.237, 0.174) and Ka was 30% higher (proportional change 0.3; 95% CI -0.71, 1.31).
Adding between-occasional variability (BOV) in the relative bioavailability significantly improved the model fit (BSV F = 15%CV, BOV F = 16%CV). In this population stavudine absorption was rapid (Ka = 10.6/h) and highly variable (BSV Ka = 142 %CV, BOV Ka = 27 %CV). The estimates of CL/F (15.2 L/h, 16 %CV) and V/F (32 L) were similar to those previously published.
Cotrimoxazole was used by 14 patients during antitubercular therapy and 10 patients continued cotrimoxazole prohylaxis after completion of antitubercular therapy.Concomitant use of cotrimoxazole increased the clearance of stavudine by 18% (95% CI 3% - 32%).
Conclusions: Rifampicin-based antitubercular therapy did not significantly affect the population pharmacokinetics of stavudine. The coadministration of cotrimoxazole and stavudine resulted in a modest increase in the apparent clearance of stavudine by 18% (95% CI 3% - 32%) but is unlikely to justify dose adjustment of stavudine when coadministered with cotrimoxazole..