Abstract

Background: Current approaches to improve the results of antileukemic treatment against AML usually involve "treatment intensification". However rational treatment intensification requires knowledge about the parameter that is to be intensified. I.e. there must exist a clearly established relation between that particular parameter and treatment effect (e.g. CR rate). A correlation of such parameters such as total drug dose, dose intensity or area under the curve (AUC) (esp. of cytarabine) and treatment effect has so far never been systematically been demonstrated in AML. Esp. AUC is often considered to best characterize an applied regimen, however it does not take into account the relative relevance of concentration versus exposure time for antileukemic efficacy. This shortcoming may be overcome by weighing the AUC by the so called concentration coefficient N which can be derived from the paradigm:

"CexpN x T = constant" for equitoxic effects (termed N-AUC).

Aims: 1) To establish the concentration coefficient N for cytarabine in AML. 2) To extrapolate from published pharmacokinetic analyses of cytarabine in AML the average AUC in each arm of a large set of randomized studies. 3) To demonstrate from these studies the existence of a linear relationship between response (CR rate) cytarabine dose, dose intensity, AUC, N-AUC.

Methods: 1) N was derived from in-vitro experiments that involved measuring equitoxic effects (LC-50%) at varying exposure times in 54 AML samples. 2) Pharmacokinetic parameters of diverse cytarabine regimens (doses ranging from 0,02 to 6,0g/m2/d; exposure times ranging from 0,5 to 24h per day) were extracted from the literature. 3) Randomized cytarabine based studies since 1990 were indentified in the literature, in which total dose, dose intensity and CR rate could be identified.

Results: 1) (Equi-)Antileukemic effects of cytarabine were well described by the paradigm

"CexpN x T = constant", which therefore made the derivation of N possible: 0,45 - N.B. that a concentration coefficient < 1 implies a higher relevance of exposure time for increased effects as compared to concentration. 2) Published cytarabine AUCs are highly linear to dose and are not affected by the infusion time, which therefore made plausible the assignment of average AUCs to the various treatment arms of other AML studies (n = 54) in which only doses and infusion times were stated. 3) No relationship could be found between CR rate and cytarabine dose (p=0,4122), dose intensity (p=0,3366) or AUC (p=0,4955). However there was a strong linear correlation between CR rate and N-AUC (r=0,5901, p<0,0001).

Conclusions: 1) We established the concentration coefficient N of cytarabine in AML, which is a useful parameter to formally quantify the relative relevance for the clinical effect of exposure time versus concentration. 2) N-AUC of cytarabine is highly correlated with the clinical effect, which suggests that it might be a better parameter for "treatment intensification" than dose, dose intensity or AUC. 3) The concept of N-AUC allows the realistic comparison of pharmacokinetically widely diverse applications of cytarabine and immediately suggests rationale ways of treatment intensification.

Author notes

Corresponding author