Predictive performance of a pharmacodynamic model for oral etoposide

The objective of this work was to prospectively validate a pharmacodynamic model for 21-day oral etoposide. The model had been developed in 27 untreated patients with stage IIIB or IV non-small cell lung cancer. Treatment consisted of 50 mg/m2/day, p.o., etoposide for 21 days in combination with 100 mg/m2, i.v., cisplatin on day 1 every 28 days for up to 6 courses. Weekly evaluations included etoposide plasma concentrations (Ec, microgram/ml) before the daily dose and WBC and neutrophil counts (ANC, 10(3)/microliters). The relationship between Ec and the pretreatment (WBCp, ANCp) and nadir counts (WBCn, ANCn) in the first course was described as follows: WBCn = 0.35 (1 + WBCp x e-1.12 x Ec)) ANCn = 0.32 (1 + ANCp x e-2.47 x Ec) The same study criteria were used to enter 26 additional patients, and 21 were evaluable for pharmacodynamics (5 had incomplete data). Predicted nadir counts were not significantly different from observed nadir counts (paired t test, P > 0.4). There were 12 and 7 patients correctly predicted to be above and below, respectively, the clinically important ANCn of 0.5 x 10(3)/microliters. The model performed reliably, and therapeutic drug monitoring appears warranted in future studies.