Prolonged exposure to topotecan in in vitro and in vivo experiments has yielded the highest antitumor efficacy. An oral formulation of topotecan with a bioavailability of 32-44% in humans enables convenient prolonged administration. Pharmacokinetic/pharmacodynamic relationships from four Phase I studies with different schedules of administration of oral topotecan in 99 adult patients with malignant solid tumors refractory to standard forms of chemotherapy were compared. Topotecan was administered as follows: (a) once daily (o.d.) for 5 days every 21 days (29 patients); (b) o.d. for 10 days every 21 days (19 patients); (c) twice daily (b.i.d.) for 10 days every 21 days (20 patients); and (d) b.i.d. for 21 days every 28 days (31 patients). Pharmacokinetic analysis was performed in 55 patients using a validated high-performance liquid chromatographic assay and noncompartmental pharmacokinetic methods. Totals of 109, 48, 64, and 59 courses were given, respectively. Dose-limiting toxicity consisted of granulocytopenia for the o.d. x 5-day dosage, a combination of myelosuppression and diarrhea in both of the 10-day schedules, and only diarrhea in the 21-day schedule. Pharmacokinetics revealed a substantial variation of the area under curve (AUC) of topotecan lactone in all of the dose schedules with a mean intrapatient variation of 25.4 +/- 31.0% (o.d. x 5), 34.5 +/- 25.0% (o.d. x 10), 96.5 +/- 70.1% (b.i.d. x 10), and 59.5 +/- 51.0% (b.i.d. x 21). Significant correlations were observed between myelotoxicity parameters and AUC(t) day 1 and AUC(t) per course of topotecan lactone. In all of the studies, similar sigmoidal relationships could be established between AUC(t) per course and the percentage decrease of WBCs. At maximum-tolerated dose level, no significant difference in AUC(t) per course was found [AUC(t) per course was 107.4 +/- 33.7 ng x h/ml (o.d. x 5), 145.3 +/- 23.8 ng x h/ml (o.d. x 10), 100.0 +/- 41.5 ng x h/ml (b.i.d. x 10), and 164.9 +/- 92.2 ng x h/ml (b.i.d. x 21), respectively.] For oral topotecan, the schedule rather than the AUC(t)-per-course seemed to be related to the type of toxicity. Prolonged oral administration resulted in intestinal side effects as a dose-limiting toxicity, and short-term administration resulted in granulocytopenia. On the basis of this pharmacokinetic study, no schedule preference could be expressed, but based on patient convenience, administration once daily for 5 days could be favored.