Low dose--high dose: what is the right dose? Pharmacokinetic modeling of etoposide.

PURPOSE

Some clinical studies on etoposide (Eto) have shown marked schedule dependency of the effect (starting at about 1 mg/l) and toxicity (over about 10 mg/l) whereas other studies have not confirmed these results. What are the conclusions we can draw from these inconsistent results when developing new low-dose (LD) and high-dose (HD) Eto schedules?

METHODS

A pharmacokinetic model for Eto based on individual pharmacokinetic data was used to simulate different LD (450 mg/m(2)) and HD (1800 mg/m(2)) schedules. The duration of exposure and the AUC of relevant concentration ranges (>1 mg/l, 1-10 mg/l, >10 mg/l) as well as peak levels were calculated in relation to the standard low dose (150 mg/m(2) over 2 h daily for 3 days).

RESULTS

The fourfold dose increase from the LD to the HD schedule was associated with a complementary increase in total AUC. However, variations in infusion time for the HD schedule were associated with large differences in AUC distribution and drug exposure with constant total AUC. Short infusions (0.5 h and 4 h) resulted in extreme peak levels (factors of 17.6+/-1.5 and 8.7+/-0.4 compared to the standard LD schedule) and an AUC >10 mg/l (factors of 17.5+/-4.9 and 17.2+/-4.8, and 83+/-2.4% and 82+/-2.4% of the corresponding total AUC), and the 96-h infusion yielded a long duration of exposure to concentrations >10 mg/l (factor 7.9+/-2.6), whereas continuous i.v. infusion over 55+/-11 days was associated with a multiple increase in the duration of exposure to "standard" drug concentrations (1 mg/l ).

CONCLUSIONS

According to evidence against schedule dependency, only target dose and pharmacokinetic variability would be appropriate rationales for Eto dosing. However, arguing for schedule dependency, simulated pharmacokinetic profiles for new Eto schedules over a wide range support that Eto-containing regimens should be designed on the basis of clear pharmacokinetic hypotheses of target levels, exposure times and AUC distributions, to allow subsequent development of pharmacokinetic/pharmacodynamic modeling.