Estimation of plasma area under the curve for etanidazole (SR 2508) in toxicity prediction and dose adjustment

The hydrophilic 2-nitroimidazole radiosensitizer etanidazole is currently undergoing clinical evaluation. Although considerably less neurotoxic than misonidazole because of its rapid renal clearance and partial exclusion from the nervous system, total dose is limited by peripheral neuropathy. Monitoring plasma etanidazole concentration in patients to determine the area under the curve (AUC0-infinity) has been proposed as a method of predicting patients at risk, and of providing a quantitative basis for dose reduction in such patients. Successful application of this policy requires accurate assessment of AUC0-infinity. We have analyzed plasma data for 18 patients receiving 2 g/m2 etanidazole to determine the errors introduced in the estimation of AUC0-infinity caused by omitting selected time points from the analysis. A 'baseline' AUC0-infinity value was calculated by integration of the rate equation for the 2-compartment model using data points at 0, 15, and 30 min and 1, 2, 4, 8, 12, and 24 hr after the end of infusion. The mean +/- SD area for AUC0-infinity was 502 +/- 152 micrograms ml-1 h (2.35 +/- 0.71 mM.h). Omitting the zero or the 24 hr time point, the average errors were quite small (2.5% in both cases), but errors of up to 16.4 and 7.3%, respectively, were seen for individual patients. Leaving out both the 8 hr and 12 hr points at the same time gave a similar low average error of 2.9%, with a highest error of 7.3%. Omitting all data points after 4 hr, the mean error was 24.7% and 15 of 18 patients had errors in excess of 10%. In addition, failure to correct for infusion time results in an underestimation of AUC0-infinity averaging 4.5% (range 1.9-8.7%). The choice of sampling times for toxicological monitoring will depend upon the accuracy with which the AUC0-infinity must be known. Including all data points between 0 and 24 hr will minimize errors. Considering the general similarity in the errors introduced by omitting the 8 hr and 12 hr points together compared to those seen with exclusion of the single 24 hr point, the choice between these truncated sampling options would be expected to lie in the relative inconvenience caused to patients and medical staff for the particular dose schedule used. The short sampling schedule (0-4 hr) should not be used.