Indirect-response model for the time course of leukopenia with anticancer drugs*

Abstract

Because both the nadir count and the duration of leukopenia after chemotherapy with anticancer drugs are important, a pharmacodynamic model describing the entire time course of leukopenia is valuable. In this study, a pharmacodynamic model was developed to simulate leukopenia.The model was developed with the 3-hour infusion data of paclitaxel. A concentration-time curve of paclitaxel for each patient estimated by a 3-compartment pharmacokinetic model was used as input to the model, which had 2 compartments corresponding to leukocytes in bone marrow and peripheral blood, respectively. Differentiation stages of myeloid cells sensitive to anticancer drugs were assumed, and exposure to a drug during their sensitive period as a function of time was used to inhibit the production of leukocytes in bone marrow. The model was validated by fitting the data of 24-hour infusion of paclitaxel or 14-day infusion of etoposide.Successful fitting of the leukopenia after a 3-hour infusion of paclitaxel was achieved. The following parameters were estimated: lag-time, 58 +/- 38 (mean +/- SD) hours before the leukocyte count started to decline; exposure giving 50% inhibition of leukocyte production (IC), 12.1 +/- 6.1 microg x h/mL; and sensitive period, 288 +/- 64 hours. These estimations were within physiologic ranges. In validation, leukopenia after 24-hour infusion of paclitaxel or 14-day infusion of etoposide was also explained by the model. Age was significantly negatively correlated with IC of paclitaxel (P = .039).This mechanistic model describes the time course of leukopenia and may provide a platform for pharmacodynamic analysis of anticancer drugs.