We aimed to evaluate the pharmacokinetics and pharmacodynamics of etoposide given chronically by the p.o. route to patients with small cell and non-small cell lung cancer. Single daily p.o. doses of 100 mg etoposide were given for 21 consecutive days every 4 weeks to 39 previously untreated patients with small cell lung cancer and 10 patients with non-small cell lung cancer. Bioavailability was studied after one i.v. and one p.o. dose of 100 mg etoposide given 48 h before and on day 1 of treatment, respectively. Etoposide plasma levels were measured using the HPLC method. Inter- and intrapatient variability of the area under the curve of the concentration versus time (AUC) during the first cycle were evaluated using a limited sampling model; the variability of etoposide plasma concentrations (Ecs) during the first cycle was assessed by weekly blood samples taken 24 h after dosing. The overall bioavailability of etoposide (mean +/- SD) was 67% +/- 22% and was not affected by fasting. A much higher inter- than intrapatient variability of both the AUC and 24-h Ec determined on days 8, 15, and 22 was found. Neutropenia was dose limiting and of varying degrees (mean +/- SD of absolute neutrophil count nadir at the first cycle: 1.5 +/- 1.2 x 10(3)/microliter). Neutropenia WHO grade >/=3 occurred in 38% of the patients after the first cycle. Pharmacodynamic analyses showed a significant relationship between the mean 24-h Ec and neutropenia, expressed as log- of absolute neutrophil count nadir or as a relative decrease of neutrophils. A correlation between a critical value of mean 24-h Ec (0.34 microgram/ml) and a high probability of achieving a greater than 80% decrease in absolute neutrophil count was found. Two pharmacodynamic models (one previously described and one developed in this study) were used to evaluate the possibility of predicting neutropenia on the basis of individual etoposide pharmacokinetics and baseline absolute neutrophil count. Pharmacokinetic studies have shown a high interpatient variability and a relatively low intrapatient variability of AUC and 24-h Ec. The application of the pharmacodynamic models and mean 24-h Ec cutoff values has proven statistically valid to predict the occurrence of severe neutropenia. However, it remains to be demonstrated in a prospective manner whether the application of pharmacokinetic/ pharmacodynamic knowledge can improve the overall therapeutic outcome of chronic p.o. treatment with etoposide.