BACKGROUND: Whether HIV infection adversely affects exposure to first-line TB drugs in children is debatable. It is also not known whether HIV infection increases the risk of plasma underexposure or overexposure to TB drugs. This study sought to address these questions.DESIGN/METHODS: Children on TB treatment were enrolled. After 4 weeks on therapy, blood samples were collected at pre-dose, 1, 2, 4, 8, and 12 h post-dose for pharmacokinetic analysis. Plasma drug exposure below and above the lower and upper bounds of the 95% confidence intervals of the reference mean for children were considered underexposure and overexposure, respectively. The effect of HIV infection on drugs exposure and risk of underexposure were examined using multivariate analysis.RESULTS: Of 86 participants (median age: 4.9 years), 45 had HIV coinfection. HIV coinfection was associated with lower pyrazinamide (PZA) and ethambutol exposures in adjusted analysis. Patients with TB-HIV coinfection were three times more likely to have PZA underexposure than those with TB only. Underexposure of rifampin was common irrespective of HIV coinfection status.CONCLUSIONS: HIV coinfection was associated with a higher risk for PZA underexposure in children. This effect should be accounted for in models and simulations to determine optimal PZA dose for children.