Pharmacokinetically guided dosing was performed in nine paediatric patients receiving etoposide. Doses on day 2 of a 2- or 3-day schedule were adapted on the basis of the day-1 area under the plasma etoposide concentration vs time curve (AUC). The day-1 AUC was estimated using a limited sampling model and the day-2 target AUC defined by the etoposide dose-AUC relationship observed in 33 children. Target AUC values (4.6-8.2 mg ml(-1) x min) were achieved with a high degree of precision and with little bias (mean error 11% and root mean squared error 15% respectively). Pharmacokinetic parameters were similar to those reported previously in children, although interpatient pharmacokinetic variability was less than that observed previously: plasma clearance, 23 (18-26) ml min(-1) m(-2); volume of distribution at steady state (Vdss), 6.0 (3.9-8.9) l m(-2); t(1/2) 254 (127-550) min (median and range). This study has demonstrated that pharmacokinetically guided dosing with etoposide is feasible. However, pharmacokinetically guided dosing is likely to be of most benefit in patients with abnormalities of renal or hepatic function, or in children with prior exposure to cisplatin.