The pharmacokinetics of oral zidovudine in HIV-infected children and adults are reported. Fourty-six patients were investigated. For data analysis three groups of similar size were formed: young children 4 months-4 years, n = 15 (group 1), older children up to 13 years, n = 16 (group 2) and young adults, n = 15 (group 3). After a single oral dose repeated blood samples were taken 1/2 hourly during a period of 4 hours and zidovudine concentrations in plasma were determined by high performance liquid chromatography. For better comparison of dose dependent parameters peak concentrations (Cmax) and the area under the time-concentration curves (AUC) were normalized either to the dose/body weight (bw) or the dose/body surface area (bs), respectively. Time to reach peak concentrations and mean terminal elimination half-life times (t1/2 beta = 63.4 +/- 47.6, 74.9 +/- 54.9 and 56.9 +/- 16.4 min in group 1, 2 and 3, respectively, mean +/- SD) were not significantly different between the three groups. With normalization to dose/bw young children in comparison to adults had significantly lower Cmax (2.7 +/- 1.3 vs. 4.6 +/- 2.4 mumol/l, p = 0.016) and AUC (226 +/- 108 vs. 373 +/- 224 mumol.min/l, p = 0.038). Group 2 gave intermediate values. However, with normalization to dose/bs differences in Cmax (6.5 +/- 3.3, 7.3 +/- 4.2 and 6.8 +/- 3.6 mumol/l, in group 1, 2, and 3, respectively) and AUC (563 +/- 313, 691 +/- 351 and 555 +/- 342 mumol.min/l, in group 1, 2 and 3) were not significant between the three groups. It is likely that changes in body water content with age may account for most of these differences observed. In conclusion, a similar pharmacokinetic profile was found in children older than 3 months as compared to older children or adults.