Serdjebi Cindy, Gattacceca Florence, Seitz Jean-François, Fein Francine, Gagnière Johan, François Eric, Abakar-Mahamat Abakar, Deplanque Gael, Rachid Madani, Lacarelle Bruno, Ciccolini Joseph, Dahan Laetitia
*SMARTc, CRO2, UMR S_911, Aix-Marseille Université, Marseille, France; †IRCM, Institut de Recherche en Cancérologie de Montpellier, INSERM U1194, ICM, Université de Montpellier, Montpellier, France; ‡Digestive Oncology Unit, La Timone Hospital, AP-HM, Aix-Marseille Université, Marseille, France; §Gastroenterology Department, CHRU Besançon, Besançon, France; ¶Digestive Pathologies Department, CHU Clermont-Ferrand, Clermont-Ferrand, France; ‖Pôle de Médecine, Centre Antoine Lacassagne, Nice, France; **Institut Arnault Tzanck, Saint Laurent du Var, France; and ††Service Interdisciplinaire de Cancérologie, Riviera-Chablais Hospital, Vevey, Switzerland.
Ther Drug Monit. 2017 Jun;39(3):290-296. doi: 10.1097/FTD.0000000000000399.
Gemcitabine remains a pillar in pancreatic cancer treatment. However, toxicities are frequently observed. Dose adjustment based on therapeutic drug monitoring might help decrease the occurrence of toxicities. In this context, this work aims at describing the pharmacokinetics (PK) of gemcitabine and its metabolite dFdU in pancreatic cancer patients and at identifying the main sources of their PK variability using a population PK approach, despite a sparse sampled-population and heterogeneous administration and sampling protocols.
Data from 38 patients were included in the analysis. The 3 optimal sampling times were determined using KineticPro and the population PK analysis was performed on Monolix. Available patient characteristics, including cytidine deaminase (CDA) status, were tested as covariates. Correlation between PK parameters and occurrence of severe hematological toxicities was also investigated.
A two-compartment model best fitted the gemcitabine and dFdU PK data (volume of distribution and clearance for gemcitabine: V1 = 45 L and CL1 = 4.03 L/min; for dFdU: V2 = 36 L and CL2 = 0.226 L/min). Renal function was found to influence gemcitabine clearance, and body surface area to impact the volume of distribution of dFdU. However, neither CDA status nor the occurrence of toxicities was correlated to PK parameters.
Despite sparse sampling and heterogeneous administration and sampling protocols, population and individual PK parameters of gemcitabine and dFdU were successfully estimated using Monolix population PK software. The estimated parameters were consistent with previously published results. Surprisingly, CDA activity did not influence gemcitabine PK, which was explained by the absence of CDA-deficient patients enrolled in the study. This work suggests that even sparse data are valuable to estimate population and individual PK parameters in patients, which will be usable to individualize the dose for an optimized benefit to risk ratio.
吉西他滨仍然是胰腺癌治疗的支柱药物。然而，毒性反应却经常出现。基于治疗药物监测进行剂量调整可能有助于减少毒性反应的发生。在此背景下，本研究旨在描述吉西他滨及其代谢产物2′,2′-二氟脱氧尿苷（dFdU）在胰腺癌患者中的药代动力学（PK）特征，并采用群体PK方法确定其PK变异性的主要来源，尽管该研究样本量少且给药和采样方案存在异质性。
分析纳入了38例患者的数据。使用KineticPro软件确定了3个最佳采样时间，并在Monolix软件上进行群体PK分析。将可用的患者特征，包括胞苷脱氨酶（CDA）状态作为协变量进行检验。还研究了PK参数与严重血液学毒性反应发生之间的相关性。
二室模型最能拟合吉西他滨和dFdU的PK数据（吉西他滨的分布容积和清除率：V1 = 45 L，CL1 = 4.03 L/min；dFdU的分布容积和清除率：V2 = 36 L，CL2 = 0.226 L/min）。发现肾功能会影响吉西他滨的清除率，而体表面积会影响dFdU的分布容积。然而，CDA状态和毒性反应的发生均与PK参数无关。
尽管采样稀疏且给药和采样方案存在异质性，但使用Monolix群体PK软件成功估算了吉西他滨和dFdU的群体和个体PK参数。估算的参数与先前发表的结果一致。令人惊讶的是，CDA活性并未影响吉西他滨的PK，这可以通过该研究中未纳入CDA缺陷患者来解释。这项研究表明，即使是稀疏的数据对于估算患者的群体和个体PK参数也很有价值，这些参数可用于个体化给药以优化获益风险比。
