The objective of this clinical and pharmacological study was to determine whether any pretreatment parameters were associated with pharmacological or toxicity parameters after prolonged oral etoposide. Therefore, the relationships between patient characteristics and etoposide concentrations and hematological toxicity were evaluated. Sixty patients with advanced non-small cell lung cancer were treated with etoposide 50 mg/m2/day p.o. for 21 consecutive days and cisplatin 100 mg/m2i.v. on day 1. Complete blood counts and etoposide plasma concentrations were obtained weekly. Etoposide was measured by high-performance liquid chromatog-raphy. The input variables were age, gender, race, weight, weight0.66. weight0.75, height, body surface area, performance status, albumin concentration, and total etoposide dose. The outcome measures were etoposide concentration; nadir values (white blood cells, neutrophils, hemoglobin, and platelets); the absolute decrease, relative decrease, and survival fraction of blood cells; and graded toxicity. No significant correlations were found in 49 fully evaluable patients between any of the input. and outcome variables. Among the outcome variables, significant correlations were found between etoposide concentration and the logarithmic transformation of the nadir blood counts. If any of the input variables were significantly correlated to etoposide concentrations or toxicity variables, it would be possible to suggest another predictor variable besides body surface area. As long as treatment is not modified for etoposide concentrations, dosing of oral etoposide must still rely on estimates of body surface area.