某醫學中心之warfarin臨床藥品動力學研究

Evaluation of Warfarin Clinical Pharmacokinetics in a Medical Center

曾碧萊(Pi-Lai Tseng)

指導教授 ： 李勇進

高雄醫學大學/藥學院/藥學研究所碩士在職專班/碩士(2007年)

摘要

目的：評估兩種不同的藥物動力學軟體 JPKD及 tdm for R，其預測warfarin劑量與INR值間準確性的比較，並探討warfarin使用的劑量及INR值的變化。
方法：以某醫學中心94年申報至健保局的資料庫中曾經使用抗凝血劑
warfarin者為研究對象。排除不符條件的病患 (如: 具有一、二、三級藥物交互作用、頻次不符、劑量未作更動等)；進一步紀錄並篩選出有作劑量更換且有已達穩定期之相關INR值者。將患者第一次的warfarin劑量及相關INR值分別輸入兩種藥物動力學軟體 JPKD及tdm for R，估算出患者個人的藥動學及藥效學參數，再輸入第二次的warfarin劑量，由軟體依上述之參數預估出更改劑量後的INR值，計算兩種軟體的絕對預測誤差APE (%) 及相似度 (Similarity)，以比較兩者的估算能力。同時分析warfarin劑量的使用情形及INR值的變化。
結果：70位未併用任何交互作用藥物的患者使用JPKD及tdm for R運算後，APE(%) 及Similarity分別為 15.0 ± 12.8、14.0 ± 10.7及0.976 ± 0.196、0.975 ± 0.175，以paired sample T test分析後p value分別為0.46；0.94。62位有併用四、五級交互作用藥物的患者使用JPKD及tdm for R運算後，APE (%) 及Similarity分別為15.1 ± 10.7、12.9 ± 10.2及0.940 ± 0.176、0.942 ± 0.154，以paired sample T test分析後p value分別為0.02；0.90。未併用任何交互作用藥物與併用四、五級交互作用藥物的患者，使用JPKD運算後，APE(%) 及Similarity以independent T test分析p value分別為0.96、0.28；使用tdm for R 運算後，APE (%) 及Similarity以independent T test分析p value分別為0.53、0.26。所有132位患者以JPKD及tdm for R運算後，APE (%) 及Similarity分別為15.1 ± 11.8、13.5 ± 10.5及0.959 ± 0.187、0.959 ± 0.166，以paired sample T test分析後p value分別為0.06；0.98。所有132位患者的實際INR值為1.57 ± 0.45，以JPKD預估之INR值為1.47 ± 0.37，以 tdm for R預估之INR值為1.47 ± 0.36。整體患者每日warfarin的使用劑量平均為2.62 ± 1.18 mg。
結論：結果顯示：1. 兩種不同的藥物動力學軟體對於未併用任何交互作用藥物及有併用四、五級交互作用藥物患者的INR值之預估能力大致相當；而當使用同一套軟體作預估時，有無併用四、五級交互作用藥物並不會影響其預估INR的能力。2. 兩種藥物動力學軟體預測之INR值與實際觀察之INR值在臨床上沒有差異。3.患者所使用的warfarin劑量與文獻建議用於中國人的使用劑量相符。

關鍵字

Warfarin ； INR ； 藥物動力學 ； JPKD ； tdm for R

並列摘要

Objects：This study was to compare the predictive ability using warfarin dosage to estimate the INR with two clinical pharmacokinetic tools, the JPKD and tdm for R. And also to evaluate the warfarin dosage and INR data in Chinese patients.
Methods：The NHI reimbursement claimed data of 2005 was retrieved from a medical center, in which all the prescribing information of warfarin in the ambulatory settings was included. However, cases such as unmatched frequency, unchanged dosage, interactions with first and second degree, abNormalized INR and/or without INR were excluded. Based on this information, cases having both of the dosage change and the INR at steady state were consequently collected for further evaluation. A set of pharmacokinetic (PK) parameters observed from the first warfarin dosage and the subsequent INR of each case was estimated by both JPKD and tdm for R. Then a new INR was estimated by using the second dosage with the previously calculated PK parameters. Thus, the precisions expressed as absolute prediction error (APE %) and Similarity could be calculated from the observed and the predicted INRs. Also, drug interaction with lessen degree was taken into account to compare the differences between the two PK models.
Results：Seventy cases without any drug interactions were studied. The APE ( %) and Similarity estimated by the JPKD and tdm for R were 15.0 ± 12.8, 14.0 ± 10.7 and 0.976 ± 0.196, 0.975 ± 0.175 respectively. There was no significance differences between these two groups using paired sample t-test (p = 0.46 and 0.94). For another 62 cases with 4 to 5 degreed drug interactions, APE ( %) and Similarity estimated by JPKD and tdm for R were 15.1 ± 10.7, 12.9 ± 10.2 and 0.940 ± 0.176, 0.942 ± 0.154 respectively. There was significance differences of APE between groups (p = 0.02), but no differences of Similarity between groups (p = 0.90). In comparison with JPKD and tdm for R model, the results were shown not significant (p=0.96 vs 0.28 and p=0.53 vs 0.26) respectively while using independent t-test to test the significant difference of the APE % and Similarity between without and with drug interaction group. At last, all the cases being with and without drug interactions were pooled together and APE % and Similarity were estimated by JPKD and tdm for R as 15.1 ± 11.8, 13.5 ± 10.5 and 0.959 ± 0.187, 0.959 ± 0.166 respectively. There was no significant differences between these two groups with the paired sample t-test (p=0.06 and 0.98). The average observed INR of all the cases was 1.57 ± 0.45, and the predicted value estimated by JPKD and tdm forR were 1.47 ± 0.37 and 1.47 ± 0.36 respectively. The average warfarin daily dose was 2.62 ± 1.18 mg.
Conculsions：Firstly, The result indicated that the predictive ability using both pharmacokinetic tools were no significant differences between the groups with and without any drug interactions. No matter what there were drug-drug interactions existed or not, no influences for both tools to predict the INR were observed. Secondly, there was no clinically differences between the observed and predicted INR values estimated by both tools. The average daily dose was similar to that recommended by the literature for Chinese people.

並列關鍵字

Warfarin ； INR ； Pharmacokinetic model ； JPKD ； tdm for R

參考文獻

1.Burton ME, Shaw LM, Schentag JJ, et al. Applied Pharmacokinetics & Pharmacodynamics Principles of Therapeutic Drug Monitoring. 4th edition. Lippincott Williams & Wilkins, 2006: pp 715-751.

2.Koda-Kimble MA, Young LY, Kradjan WA, et al. Applied Therapeutics: The Clinical Use of Drugs. 8th edition Lippincott Williams & Wilkins, 2004: chapter 16.

3.Pirmohamed M, James S, Meakin S, et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18820 patients. British Medical Journal 2004;329(7456):15-9.

4.Blix HS, Viktil KK, Reikvam A, et al. The majority of hospitalised patients have drug-related problems: results from a prospective study in general hospitals. European Journal of Clinical Pharmacology 2004;60(9):651-8.

5.Buffington DE, Lampasona V, Chandler MH, et al. Computers in pharmacokinetics. Choosing software for clinical decision making. Clinical Pharmacokinetics 1993;25(3):205-16.