術前肝動脈化學栓塞的使用對可切除性肝癌病人術後醫療資源耗用與存活之研究-系統性回顧與整合分析

研究目的本研究以全國性資料庫，篩選出肝切除術(Hepatectomy)術前使用肝動脈化學栓塞(TransArterial ChemoEmbolization, TACE)的TACE組與單純使用肝切除術的LR組肝癌病患作為研究對象，另為消除選擇偏誤(Selection bias)，執行傾向分數配對(Propensity Score Matching, PSM)，探討術前TACE對台灣的可切除性肝癌患者術後療效的影響，再者本研究利用已發表文獻進行整合分析(Meta Analysis)，討論國內外的差異，以供未來臨床單位、學術單位及公共衛生單位在相關議題決策上之參考。研究方法本研究設計分兩部分。第一部分使用十年以上全國性資料庫進行探討，以可切除性肝癌病患為基礎，並區分術前TACE組和LR組，研究樣本共11,079位，其中年齡、性別、肝炎種類、肝硬化、CC指數、醫院層級、醫院/醫師手術量等變項經PSM後，TCAE組和LR組各為646人，進而分析上述變項對術後療效的影響，利用SPSS 19.0的Cox迴歸以及線性迴歸等方法。第二部分的整合分析，以PubMed、ProQuest Health & Medical、MEDLINE和Google等資料庫，搜尋1989/1/1~2012/4/25以可切除性肝癌病人為基礎，術前施行TACE為實驗組，單純施行肝切除術為對照組的文獻，一共有18篇文獻符合標準而納入分析。由兩位研究者獨立萃取研究特性與結果指標後，以隨機效應模式整合分析進行森林圖(Forest Plot)、異質性檢定(Heterogeneity Test)、敏感度分析(Sensitivity Analysis)、漏斗圖(Funnel Plot)等方法，以取得術前TACE對術後療效的肯定結論。研究結果 PSM後，TACE組患者在住院天數和醫療費用顯著較LR組患者高(49.2天vs. 19.1天；11,702元 vs. 4,625元)，但在第5年的無病存活率和整體存活率則無顯著差異。住院天數的顯著影響因子，包含肝硬化、合併症指數、醫院層級和術前TACE的執行；醫療費用的顯著影響因子，包含合併症指數和術前TACE的執行。五年無病存活率的顯著影響因子，包含肝炎、肝硬化、CCI指數、醫院層級、醫院/醫師手術量；五年整體存活率的顯著影響因子，包含肝炎、合併症指數、醫院層級、醫院手術量。整合分析發現，術前TACE組患者相對於LR組患者，五年無病存活的機率為1.39倍(95%信賴區間為1.00-1.93，P=0.054)；五年整體存活的機率為0.71倍(95%信賴區間為0.46-1.07，P=0.100)；肝外復發的機率為0.96倍(95%信賴區間為0.53-1.72，P=0.884)，皆無顯著差異。結論與建議以住院天數來看，醫師手術量較醫院手術量影響來的大，施行術前TACE組的患者顯著有較長住院天數；在醫療費用的觀察也有類似結論。術前TACE是一個相當安全的術式，對於院內死亡無顯著影響，但對於五年無病存活率和五年整體存活率亦無顯著影響。本研究在研究上仍有限制，因本研究無法得知如腫瘤分期與腫瘤組織學檢查等臨床變項，可能對於本研究的結果解釋有所影響；在整合分析方面，因文獻數量的限制，在納入/排除標準無法更嚴格定義。另外，本研究在術前TACE與純肝切除術的兩組病患比較上，是第一篇以全國性資料庫進行分析且執行PSM的研究，同時也是第一篇不以隨機分配文獻為限的整合分析，在研究結果的推估上較以往文獻有說服力。由於執行術前TACE有顯著較多的住院天數和醫療花費，但對於術後存活並無顯著影響，因此本研究建議，可切除性肝癌病人應盡量避免術前TACE的施行。

關鍵字

術前肝動脈化學栓塞；全國性資料庫回溯研究；傾向分數配對；整合分析

並列摘要

Objective There is heterogeneity still in the effect of preoperative transArterial chemoembolization(TACE) for the outcome of the resectable hepatocellular carcinoma(HCC) patients. The main aims of this research are to investigate the influence of preoperative TACE for resectable HCC patient's outcome. Methods This study is divided into two parts. First, this research retrospectively the patients’ variables, such as age, gender, hepatitis, cirrhosis, comorbidity index(CCI), hospital level, hospital/surgical volume from the national databases for over 10 years. In addition, this research solving the problems of the selection bias of the samples by propensity score matching(PSM). In the last, the patients who under preoperative TACE(TACE group, n = 646)and those who did not(LR group, n = 646) was involved in this study, using Logistic regression and Cox regression of SPSS 19.0. Second, this study integrated the published literature to meta analysis. There are 18 published literatures were involved. Data were synthesized using the random effects models. Results were expressed as hazard ratio(HR) with 95% confidence intervals. Results After PSM, The TACE group’s length of stay(LOS) and resource utilization are significantly higher than LR group’s(49.2 days versus 19.1 days;$11,702 versus $4,625), but the difference of 5 years disease-free survival and(DFS) overall survival(OS) between the two groups aren’t significantly. The LOS significant impact factors include cirrhosis, CCI, hospital level and Preoperative TACE using; The resource utilization significant impact factors include CCI and Preoperative TACE using. The 5-year DFS significant impact factors include hepatitis, cirrhosis, CCI, hospital level, hospital/surgical volume; The 5-year OS significant impact factors include hepatitis, CCI, hospital level, hospital volume. In the meta analysis, the 5-year DFS of the TACE group (odds ratio = 1.39, 95% CI 1.00-1.93; P = 0.054), the 5-year OS of the TACE group (odds ratio = 0.71, 95% CI 0.46-1.07; P = 0.100) and the recurrence of theTACE group (odds ratio = 0.96, 95% CI 0.53-1.72; P = 0.884). Conclusion and suggestion Preoperative TACE is safe procedure, there are not signicant in-hospital mortality and postoperative mortality(within 1 month) betwwen two groups. The using of preoperative TACE is linked with higher LOS and resource utilization, but it isn’t significant influent in 5-year DFS and 5-year OS. The limitations of the study invole that the national database can’t show the clinical variables, and the published literatures are not enough to make the criteria of the selection seriously. In the conclusion, the resectable HCC patients who use preoperative TACE will lead to higher LOS and resource utilization, but it’s not influence in promoting 5-years DFS and 5-years OS. The research suggested that the resectable HCC patients should avoid using preoperative TACE.