腎功能不全病人之Warfarin群體藥動學與劑量算則

Abstract

研究背景 Warfarin為一抗凝血劑，係由R-及S-warfarin鏡像異構物消旋組成。Warfarin 主要經肝臟代謝，因而仿單標示，使用於腎功能不全病人時無需特別做劑量的調整。然而，近年來有文獻顯示高加索人、非裔美國人及日本人之腎功能狀態與warfarin用藥劑量有著顯著的相關性。 研究目的 本研究探討臺灣人之S-warfarin (S-warfarin之抗凝血作用為R-warfarin的2-5倍) 群體藥動學參數，是否受腎功能狀態影響，並嘗試建立腎功能不全病人的warfarin劑量算則。 研究方法 本研究納入對象為使用warfarin且劑量穩定達兩周以上之心臟外科、心臟內科及抗凝血門診病人，以HPLC-UV方法定量S-warfarin的血漿濃度，採PCR-RFLP方法檢測CYP2C9*3 與VKORC1-1639 基因型。血清白蛋白、尿素氮、C-反應蛋白、肌酸酐、胱蛋白C使用光譜測定法。群體藥動學模型建立採非線性混合效應模式，使用軟體為Monolix，warfarin劑量算則運用多元線性回歸法建立，使用軟體為R。 研究結果 本研究共納入75位warfarin用藥病人，平均年齡為63.6 ± 9.2 歲，男女比約2:1。平均肌酸酐清除率為57.8 ± 21.4 mL/min，warfarin平均日劑量為2.7 ± 1.1 mg。本研究顯示肌酸酐清除率與warfarin平均日劑量呈現顯著相關性 (R2 = 0.18，p < 0.001)。在群體藥動學PPK的分析上，發現肌酸酐清除率與benzbromarone併用 (CYP2C9抑制劑) 顯著影響S-warfarin的清除率;胱蛋白C與benzbromarone併用則對S-warfarin的分佈體積有顯著影響(R2 = 0.2806, p < 0.001)。在進一步排除併用benzbromarone 的病人之次族群分析中，肌酸酐清除率、amiodarone併用 ( CYP2C9 與CYP3A4 抑制劑)及 CYP2C9*3 基因型，為S-warfarin清除率顯著影響因子;而胱蛋白C則為影響S-warfarin分佈體積的顯著因子 (R2 = 0.2852, p < 0.001)。在多元線性回歸分析後，發現胱蛋白C、INR、CYP2C9*3與VKORC1-1639 基因型，及 amiodarone併用，同為warfarin劑量算則的顯著因子。 結論 腎功能狀態與warfarin日劑量、清除率與分佈體積具有高度相關性。本研究係首先導入尿毒素之一的胱蛋白C於S-warfarin群體藥動學之分佈體積模型與warfarin劑量算則之建立。本研究指出warfarin用於腎功能不全病人時。劑量調整是必要的。 關鍵字: warfarin、胱蛋白C、腎功能不全、群體藥動學、劑量算則

Background The anticoagulant warfarin, a racemic mixture of R- and S-enantiomers, is metabolized mainly in the liver and bears a label of dose adjustment not necessary for patients with renal impairment. However, recent studies revealed an association between warfarin dosage and renal function status in Caucasians, African Americans, and Japanese. Objectives The study was aimed to investigate the impacts of renal function status on the population pharmacokinetics of the index enantiomer S-form, 2 to 5 times more potent than the R-form, and to develop warfarin dosing algorithm for patients with renal impairment in Taiwanese. Methods Patients who were on warfarin and under constant doses for at least two weeks were recruited from Cardiology clinics (surgical and medical) and the pharmacist-run anticoagulant clinic. Plasma S-warfarin levels were measured by HPLC-UV and genotypes of CYP2C9*3 and VKORC1-1639 were determined by PCR-RFLP. Serum albumin, blood urea nitrogen (BUN), C-reaction protein (CRP), serum creatinine (sCr), and cystatin C (CysC) were measured by spectrophotometric analyses. Population pharmacokinetic parameters was estimated by nonlinear mixed effect modeling using Monolix. Dosing algorithms were established by multivariate linear regression using R. Results A total of 75 patients, with a mean age of 63.6 ± 9.2 years old and male-to-female ratio of 2:1, were recruited in the study. The mean creatinine clearance (CrCl) and warfarin daily dose were 57.8 ± 21.4 mL/min and 2.7 ± 1.1 mg, respectively. An association between warfarin daily dose and CrCl was observed (R2 = 0.18, p < 0.001). Population pharmacokinetic analyses demonstrated that CrCl and benzbromarone (a CYP2C9 inhibitor) comedication contributed significantly to the clearance (Cl) of S-warfarin. Volume of distribution (Vd) of S-warfarin was significantly affected by CysC and benzbromarone comedication (R2 = 0.2806, p < 0.001). By excluding patients comedicated with benzbromarone, the subgroup analysis showed CrCl, amiodarone (a CYP2C9 and CYP3A4 inhibitor) comedication, and CYP2C9*3 genotype were significant covariates of the Cl of S-warfarin, whereas CysC was a significant covariate for the Vd of S-warfarin (R2 = 0.2852, p < 0.001). Further multivariate linear regression analysis showed that CysC, INR, CYP2C9*3 and VKORC1-1639 genotypes, and amiodarone comedication were significant covariates in describing warfarin dosing algorithm. Conclusion Renal function is highly correlated with the daily dose, Cl, and Vd in warfarin users with renal impairment. The incorporation of a uremic toxin, CysC, into both the S-warfarin Vd models and warfarin dosing algorithm is also uniquely presented. Dose adjustment in warfarin prescriptions is definitely recommended for patients with renal impairment. Key words: warfarin, cystatin C, renal impairment, population pharmacokinetics, dosing algorithm