- 學位論文
利用資料蒐納方法探討Digoxin在台灣之族群藥物動力學研究
The Population Pharmacokinetics of Digoxin in Taiwan Study Using Data Mining Technique
摘要
本研究之目的為利用資料蒐納的方法，探討影響在台灣使用毛地黃(Digoxin)的病患族群其血中濃度的顯著相關因素，當鑑別出此影響因素後便可當作族群藥物動力學研究的共變相，繼而導出族群藥物動力學方程式。以研究出的危險因素可作為醫院在給予Digoxin治療時之警訊工具，幫助醫師在評估Digoxin之臨床治療效果。 研究中以回溯性方法收集118個單點抽血資料當作資料蒐納背景，病患分別以Digoxin血中濃度，疾病狀態(鬱血性心臟衰竭、心房纖維顫動、竇性心跳過速)，生理狀態(腎功能、肝功能)，藥物與藥物交互作用(鈣離子阻斷劑、血管緊縮素轉化酵素抑制劑、制酸劑、利尿劑)及其他可能影響的因素(如性別、體重、血壓、年齡及電解質)為項次來記錄資料蒐納分析研究。利用PMLSF(C4.5)的資料蒐納技術來預測臨床上這些分類項次與Digoxin血中濃度的相關性，而決定出顯著的影響因素，以十倍的交叉確效方法來確認分類的準確性。而在族群藥物動力學之分析屬性包括性別、體重、肌酸酐廓清率、鬱血性心臟衰竭，以58位合於納入條件之成人抽血資料為樣本，利用非線性混合效應模式(NONMEM)軟體估算族群藥物動力學參數，採用ADVAN2，TRANS2模式計算族群藥物動力學方程式。 由結果顯示: PMLSF(C4.5)以十倍的交叉確效方法來確認分類的準確率達64.088%，其中體重、鉀離子濃度、鬱血性心臟衰竭、年齡、性別、肌酸酐廓清率、每日使用劑量、心房纖維顫動、血壓、竇性心跳過速及鈉離子分別為影響Digoxin血中濃度的重要因子。 本研究再由資料蒐納分析結果選擇了性別、年齡、體重、肌酸酐廓清率及鬱血性心臟衰竭當作NONMEM的屬性進行估算族群藥物動力學參數，在最小觀察值為40.552之最適化的條件下，求得Digoxin平均廓清率CL=(0.111´CLcr+1.72´CHF+3.44´SEX)´(1.306)。由所估算的族群藥物動力學參數可以幫助臨床醫師作個別病患使用Digoxin的藥物投予設計，本研究亦建立藥物動力學參數與地區因素及個別病患間共變數之相關性。
並列摘要
The purpose of this study was to explore the relative and clinical factors by Data Mining that influenced digoxin therapeutic levels in Taiwanese. After these factors were identified, they would be used as covariates in digoxin population pharmacokinetic study and estimated the equation with retrospective cases. The risk factors will be an implement as the alarm system to assist physician for digoxin therapy. In this study, one hundred and eighty one single sampling data were collected retrospectively to be the data mining background. Patients demographic data were recorded as well as digoxin serum levels, disease factors (such as congestive heart failure, arterial fibrillation, and sinus tachycardia), physiological status (renal and liver function), drug-drug interactions (calcium channel blocker, angiotensin conversing enzyme inhibitors, antacid, and diuretics), and other possible factors would also be included (such as sex, age, body weight, blood pressure, and electrolyte). The data mining algorithms PMLSF(C4.5) were adopted to train classification data, and the classification scheme would be tested by 10-fold cross-over validation method to validate the accuracy of classification method. The study of population pharmacokinetics was attributed to the record as well as sex, age, body weight, serum creatinine clearance and congestive heart failure. Fifty-eight adult patients satisfied with the inclusive criteria were collected retrospectively and went forward to the population pharmacokinetic analysis. The NONMEM was adopted to estimate population pharmacokinetic parameters. The estimation model was performed using ADVAN 2 and TRANS 2 to calculate the pharmacokinetic equation. The results showed that PMLSF(C4.5) correctly classified rate from 10-fold cross-over validation procedure was 64.0884%. In PMLSF, body weight, potassium level, congestive heart failure, age, sex, serum creatinine clearance, daily dose, atrial fibrillation, blood pressure, sinus tachycardia, and sodium level were selected as key factors out of selected candidate’s factors. According the information from the data mining, we chose these factors including sex, age, body weight, serum creatinine clearance, and congestive heart failure to be the attribution to estimate by NONMEM. The final population pharmacokinetic equation for Digoxin clearance was CL=(0.111´CLcr+1.72´CHF+3.44´SEX)´(1.306), and the minimum value of objective function was 40.552. The estimated values of population parameters may assist clinicians in the individualization of digoxin dosage regimens. The relationships among the pharmacokinetic parameters, the demographic data, and the patient-specific covariates were established in this study.
參考文獻
延伸閱讀
- 黃翠娥（2018）。野上彌生子的台灣觀點研究─以〈台灣〉為探討中心─｜Discussion of Yaeko Nogami's perspectives on Taiwan-Based on the Taiwan。台大日本語文研究，(36)，47-68。https://doi.org/10.6183/NTUJP.2018.36.47
- 林俊清、顏銘宏（1992）。台灣產生藥資源之開發研究（第九報）登豎杇之生藥學研究。中醫藥雜誌，2(2)，33-49。https://doi.org/10.6940/JCM.199204_2(2).04
- 張容榕（2006）。建立台灣地區流行性嗜血桿菌分子指紋圖譜資料庫及其抗藥性相關之研究〔碩士論文，臺北醫學大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0007-1704200715034701
- 林建成（2012）。Using Data Mining Techniques for the Clustering of Elderly People in Taiwan And Analyzing Their Relationships with Life Satisfaction and Life Model Factors〔碩士論文，長榮大學〕。華藝線上圖書館。https://doi.org/10.6833/CJCU.2012.00098
- 黃冠豪（2017）。A Research of Applying Data Mining Techniques to Explore the Associated Rules to Department Application and Entrance of Fresh Newcomers in Taiwanese Universities〔碩士論文，義守大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0074-1907201714134100