BACKGROUND: In 1986, the concept of pharmacokinetically guided dose escalation (PGDE) was proposed to predict the maximum tolerated dose (MTD) of an antitumor drug in humans from animal data. We have previously shown that antitumor drugs can be classified into two types, depending on their cytotoxic mechanisms: type 1 drugs, which are cell cycle phase-nonspecific agents, i.e., area under the curve for drug concentration in the plasma versus time (AUC)-dependent drugs; and type 2 drugs, which are cell cycle phase-specific agents, i.e., those that are time dependent. PURPOSE: The validity of the assumption that the AUC at the dose lethal for 10% of mice administered drug (LD10) is equal to the AUC at MTD for humans, the premise on which PGDE is based, was examined for type 1 and 2 drugs. METHODS: Findings in the literature, including those of Collins and co-workers, were retrospectively analyzed. The human/mouse ratios for the AUC were compared with each other and with the human/mouse dose ratios, based on milligram per meter square of body surface area, the measurement currently used in clinical trials of antitumor drugs. For six of the type 1 drugs, the human/mouse ratio for the AUC of total drug (AUC) and that of unbound drug (AUCu), which has been considered a determinant of pharmacologic and toxicologic effects, were also compared. RESULTS: There was an excellent correlation between log AUC at LD10 for mice and log AUC at MTD for humans for type 1 drugs (r = .898), but not for type 2 drugs (r = .677). For type 1 drugs, the correlation between mouse AUC at LD10 and human AUC at MTD was better for unbound drug (r = .961) than for total drug (r = .892). CONCLUSIONS: PGDE is useful for type 1 drugs; differences in protein binding between species should, however, be considered when using this method.
BACKGROUND: In 1986, the concept of pharmacokinetically guided dose escalation (PGDE) was proposed to predict the maximum tolerated dose (MTD) of an antitumor drug in humans from animal data. We have previously shown that antitumor drugs can be classified into two types, depending on their cytotoxic mechanisms: type 1 drugs, which are cell cycle phase-nonspecific agents, i.e., area under the curve for drug concentration in the plasma versus time (AUC)-dependent drugs; and type 2 drugs, which are cell cycle phase-specific agents, i.e., those that are time dependent. PURPOSE: The validity of the assumption that the AUC at the dose lethal for 10% of mice administered drug (LD10) is equal to the AUC at MTD for humans, the premise on which PGDE is based, was examined for type 1 and 2 drugs. METHODS: Findings in the literature, including those of Collins and co-workers, were retrospectively analyzed. The human/mouse ratios for the AUC were compared with each other and with the human/mouse dose ratios, based on milligram per meter square of body surface area, the measurement currently used in clinical trials of antitumor drugs. For six of the type 1 drugs, the human/mouse ratio for the AUC of total drug (AUC) and that of unbound drug (AUCu), which has been considered a determinant of pharmacologic and toxicologic effects, were also compared. RESULTS: There was an excellent correlation between log AUC at LD10 for mice and log AUC at MTD for humans for type 1 drugs (r = .898), but not for type 2 drugs (r = .677). For type 1 drugs, the correlation between mouse AUC at LD10 and humanAUC at MTD was better for unbound drug (r = .961) than for total drug (r = .892). CONCLUSIONS:PGDE is useful for type 1 drugs; differences in protein binding between species should, however, be considered when using this method.