OBJECTIVE To define the pharmacokinetics of isoniazid (INH) in children with tuberculosis in relation to the N-acetyltransferase 2 (NAT2) genotype. METHODS The first order elimination rate constant (k) and area under the concentration curve (AUC) were calculated in 64 children <13 years of age (median 3.8) with respiratory tuberculosis from INH concentrations determined 2-5 hours after a 10 mg/kg INH dose. The NAT2 genotype was determined; 25 children were classified as homozygous slow (SS), 24 as heterozygous fast (FS), and 15 as homozygous fast (FF) acetylators. RESULTS The mean (SD) k values of the genotypes differed significantly from one another: SS 0.254 (0.046), FS 0.513 (0.074), FF 0.653 (0.117). Within each genotype a median regression of k on age showed a significant decrease in k with age. The mean (SD) INH concentrations (mg/l) two hours after INH administration were SS 8.599 (1.974), FS 5.131 (1.864), and FF 3.938 (1.754). A within genotype regression of 2-hour INH concentrations on age showed a significant increase with age. A within genotype regression of 3-hour, 4-hour, and 5-hour concentrations on age also showed a significant increase with age in each instance. In ethnically similar adults, mean (SD) 2-hour INH concentrations (mg/l) for each genotype were significantly higher than the children's: SS 10.942 (1.740), FS 8.702 (1.841), and FF 6.031 (1.431). CONCLUSIONS Younger children eliminate INH faster than older children and, as a group, faster than adults, and require a higher mg/kg body weight INH dose to achieve serum concentrations comparable to adults.