Comparison between NONMEM and NPAG for gemcitabine modelling
Aldaz A. (1), Sayar O. (1), Zufia L.(1), Viudez A.(2), Rifon J.(3), Nieto Y.(4)
(1) Pharmacy Department, University Clinic of Navarra, Pamplona, Spain.(2) Oncology Department, University Clinic of Navarra, Pamplona, Spain. (3) Hematology Department, University Clinic of Navarra, Pamplona, Spain.(4) Oncology Department, M.D. Anderson Cancer Center, University of Texas, USA.
Objectives: Gemcitabine is a prodrug with proved efficacy in diferents tumors as lung or pancreas cancer. There is a good correlation between the controlled rate of infusion of this drug and the intracellulary accumulation of its active metabolites (the forms di- and triphosphate) and also between gemcitabine plasma concentrations and the level of its intracelular activation. A good pharmacokinetic model to fit gemcitabine concentrations could improve the clinical use of this drug.
Methods: The NONMEM v. 5 and the NPAG from the USCPack v11.8 was used to fit the concentrations of gemcitabine using a monocompartimental model. Eight plasma samples per patient were draw from 64 oncologic patients in treatment with gemcitabine and docetaxel. The assay error equation was 0.0441+0.0429C+0.0034C2. The differences in the means from both programs were compared by t-student test.
Results: Mean (Standard deviation) volumen of distribution estimation with NONMEM was 36.4(3.25) and with NPAG 45.34(30.39), p=0.02 for the difference. For the total clearance these values are 142.45(18.04) with NONMEM and 137.39(29.06) with NPAG, not statistically different (p=0.24). In both parameters variation of coefficients were greater with NPAG.
Conclusion: Mean gemcitabine pharmacokinetic parameters estimated by NONMEM and NPAG were similar, but their variability was greater with NPAG. Therefore it seems easier to detect cases with parameters clearly different from the mean with NPAG, and then more usefull to clinical practice.