Twenty-two pharmacokinetic studies were carried out in 11 patients receiving cisplatin (20 mg/m2 per d) associated with etoposide (50 mg/m2 per d), as 5-day continuous infusions, every 4 weeks. Blood was withdrawn at 8:30 am from day 1-5. Within 15 min after taking the blood, an aliquot of plasma was filtered for the ultrafilterable platinum (UP) assay. Total platinum (TP) and UP were assayed by flameless atomic absorption. The plasma concentrations and AUC0-120 h of TP were correlated with those of UP (P less than 0.05 to P less than 0.001). TP concentrations increased significantly during the infusion and with each successive course, whereas the increase of plasma concentration of UP during and between courses was not statistically significant. The responders had significantly higher levels of TP (AUC, concentrations) in the first and second courses than the non-responders. No renal toxicity was observed, nevertheless, the AUC0-120 h of TP and UP were positively correlated with the serum creatinine (P less than 0.05). The digestive intolerance (grade 1-3) was significantly correlated with TP concentrations and AUC0-120 h. There was no statistical difference in UP concentrations either between responders and non-responders in any course, nor between toxic and non-toxic courses. Since etoposide was concomitantly administered, we can formulate the conclusion as follows: no "objective" response was observed in the patients with low TP plasma concentrations and AUC0-120 h.