台灣單一中心肝臟移植結果與多面向因子之研究

An analysis of multidimensional factors and outcomes for liver transplantation in single center of Taiwan

謝宗保(Chung-Bao Hsieh)

指導教授 ： 鍾國彪

國立臺灣大學/公共衛生學院/健康政策與管理研究所/博士(2014年)

https://doi.org/10.6342/NTU.2014.00269

摘要

背景
肝臟移植是治療末期肝臟疾病或肝細胞癌的最後選擇。在台灣，由於腦死捐贈者極為稀少，親屬活體肝臟移植已成為一種替代且普遍的手術方式。而肝臟移植的成果已經被廣泛的討論及研究，包括在內部可以監測醫療的照護品質，外在的部分可以做為政府機構或大眾媒體監督的依據，二者均可影響健康照護體系。
目前研究的臨床結果均不夠全面性，大多只考慮受贈者風險因素：如疾病嚴重程度、肝硬化程度、或肝癌病患其癌症顆數及大小，會影響其臨床結果。然而，肝臟移植是一種極複雜的手術：其中包含了捐贈者及受贈者二部分。捐贈者風險因素、捐贈者與受贈者配對相關因素：如植入肝重量配對、性別配對、年齡配對、植入肝與疾病嚴重程度的配對等，也會影響受贈者的臨床結果。除此之外，手術醫師的技術及經驗，移植術後服用藥物、術後排斥、術後感染、術後第二腫瘤發生，皆影響肝臟移植的成果。
然而，目前並沒有任何研究可以包含上述數個面向。統計的方法也無法包含醫師技術及經驗如學習曲線,捐贈者與受贈者配對相關因素加以分析、探討其間之相互關係。本研究目的是探討上述數個面向，探討影響肝臟移植臨床結果之風險因素分析。醫師學習曲線,捐贈者與受贈者配對相關因素是否影響肝臟移植的成果
研究設計
單一中心連續性移植病人,回溯性研究
材料與方法
研究方法：本研究於單一醫學中心回溯性分析捐贈者、受贈者、捐贈者與受贈者配對性變項、移植醫師相關因素、肝臟移植術後因素與肝臟移植臨床結果之相關性。臨床結果包括受贈者的術後膽管併發症、術後30天神經併發症、受贈者術後30天及90天內的死亡與總體死亡。所有接受肝臟移植病患之資料，包含上述數個面向從第一例到第295例與臨床結果，包含手術醫師或團隊第一例到第N例以利表示學習曲線。利用R統計軟體做logistic迴歸分析及Cox’s迴歸分析以找出線性因素，以GAM(Generalized additive models)找出非線性因素含學習曲線，將此二者放入多變項分析，找出影響肝臟移植臨床結果之危險因素最適當之模式。
結果：
在受贈者的術後膽管併發症研究中發現危險因素為：受贈者面向包含共病嚴重程度、術前血氨濃度、術前肌酐酸濃度、白蛋白濃度、肝硬化嚴重程度、術中出血量、右肝移植；捐贈者與受贈者配對面向為年齡乘積、植入肝重量與肝硬化嚴重度比、植入肝重量與末期肝病嚴重度比；移植醫師面向為學習曲線前九十例及A醫師；術後併發症面向包含術後急性排斥及術後洗腎；外在環境面向為健保政後。此統計模式concordance為0.806。
在受贈者的術後神經併發症研究中發現危險因素為：受贈者面向包含年齡、身體質量指數、移植時Child score、肝昏迷病史、精神疾病、無靜脈曲張出血病史、術後七天FK濃度、腹腔內感染；捐贈者面向為年齡；配對面向包含男男配對、植入肝重量與受贈者重量比；移植醫師面向為學習曲線前三十例。此統計模式AUC為0.855。
在受贈者術後30天內死亡的危險因素為：受贈者面向為肝臟末期疾病嚴重程度>28分；配對面向為女女配對；術中面向為術中出血量；術後併發症面向為植入肝失能。此統計模式AUC為0.961。在受贈者術後90天內死亡的危險因素為：受贈者面向包含肝臟末期疾病嚴重程度>28分、術前血糖>145mg/dL；捐贈者面向為年齡；配對面向為女女配對及植入肝重量與受贈者重量比；術中面向為手術時間；術後併發症面向為植入肝失能、術後第七天FK濃度、嚴重術後併發症>IIIA；移植醫師面向為團隊移植個數介於121至246。此統計模式AUC為0.962。
在受贈者的長期總體死亡研究中發現危險因素為：受贈者面向包含年齡、飲酒史、無靜脈曲張出血史、肝臟末期疾病嚴重程度>28分、身體質量指數；捐贈者與受贈者配對面向為女性給男性、植入肝重量與受贈者重量比；術中面向為手術時間；移植醫師面向為團隊移植個數<50；術後併發症面向包含術後洗腎、腹內感染、植入肝失能、術後第七天鈉離子濃度、術後腫瘤復發、術後第二種腫瘤、術後肺結核感染、術後急性排斥。此統計模式concordance為0.890。
結論：
影響上述四項肝臟移植臨床結果，其危險因素皆為多面向，包括受贈者、捐贈者、捐贈者與受贈者配對因素、術中因素、移植醫師因素、術後併發症、外在環境變化等。在控制多變項危險因子後，團隊的學習曲線仍會影響長期存活及神經併發症，但不會影響30天及90天的死亡；移植醫師的學習曲線會影響術後膽管併發症。
貢獻性：
本研究闡明受贈者、捐贈者、捐贈者與受贈者配對因素、術中因素、移植醫師因素、術後併發症、外在環境變化等會影響肝臟移植臨床結果，這些是被以前研究所忽略。此研究結果可以幫助醫師避開危險因子，以改善臨床結果。此研究將應用R統計分析技巧以處理叢聚性資料，在以往的研究中往往忽略這個問題：此結果讓醫院管理者重視完整的訓練計畫以增加團隊成熟度及累積移植醫師的經驗，進而提升肝臟移植的臨床結果。

關鍵字

肝臟移植 ； 捐贈者與受贈者配對因素 ； 移植醫師及團隊累積個數 ； 學習曲線 ； 術後膽管併發症 ； 術後神經併發症 ； 術後早期死亡 ； 長期存活率 ； 危險因子 ； 多變項迴歸分析 ； 非線性因素分析

並列摘要

Background:
Liver transplantation (LT) becomes the final therapeutic choice for end-stage liver disease or hepatocellular carcinoma. In Taiwan, living-related donor liver transplantation (LDLT) is an alternative and become popular owe to the shortage of cadaveric donor. The outcome of LT has been studied and discussed worldwide, either monitor the quality of medical care from internally; or being the guideline of monitor by the press or government from externally. Both of them could affect the health care system.
Recent outcome studies were focused on risk factors of liver recipients, such as cormorbilities, the degree of liver cirrhosis, or patients with hepatocellular carcinoma whose tumor number and size, were influent the clinical outcome. However, LT is a very elaborate surgery comprising both the donor and recipient parts. The risk factors of liver donor, the matched factors between donor and recipient: e.g. Graft recipient weight ratio (GRWR), gender matched factor, age matched factor, are also influent the clinical outcomes. Besides, the experience and refinement techniques of transplant surgeon, the medication after liver transplant, the rejection episode, the infection disease e.g. T.B. or CMV infected and De Novo cancer were affect the clinical outcomes of LT.
Nowadays, there are no any studies have contained all dimensions and clarified the interaction between each risk factors. The statistics method which used in the previous risk factor studied could not analysis the surgeon and team learning curve, donor-recipient matched factors and non-linear variables. The purpose of this study is tried to define the risk factors to influent the clinical outcomes of LT from all dimensions aformentioned.
Materials and Methods:
In this retrospective study, we collected the data of donor, recipient, donor and recipient matched variables, transplant surgeon related factors, and post transplant related factors and determined the risk factors of clinical outcomes of liver transplant from single medical center. The clinical outcomes were defined as the post-transplant neurological complications, post-transplant biliary complications, and the early mortality in 30 days & 90 days after transplant, and overall mortality. A total of 295 patients undergoing LT at Tri-Service General Hospital were enrolled. The collected data were analyzed using the R statistic software to fit the logistic regression analysis and Cox’s regression analysis to figure out the linear factors. Non-linear factors were examined to fit the generalized additive models (GAM).All the factors were used for the multiple variables analysis and to determine the risk factors to influent the clinical outcomes of LT.
Result:
The risk factors for post-transplant neurological complications were: dimension of recipient included age <29 or >60 y/o, body mass index (BMI) <21.6 or >27.6 kg/m2, high Child-Pugh score while transplant, positive hepatic coma history, positive psychiatry disease history, negative of varices bleeding history, higher FK level in post-transplant 7 days, positive of intra-abdominal infection (IAI); the dimension of donor is donor’s age <22 or >40 y/o; the dimension of donor-recipient matched included male to male matched, graft recipient weight ratio (GRWR) between 0.9% to 1.9%; the dimension of transplant surgeon is learning curve within first 31 case. The area under curve of this statistic model was 0.855.
The risk factors for post-transplant biliary complications included dimension of recipient included Child-Pugh score between 7 to 11, Carlson index between 4 to 6 , the ammonia level <70 mg/dl in transplant, the higher serum creatinine level in transplant, the higher serum albumin level in transplant; the dimension of donor-recipient matched included the GRWR and Child score ratio <0.08 or >0.2, GRWR and MELD score ratio <0.08 or >0.2, donor age cross recipient age >2000; the dimension of operation included blood loss >2800 ml, partial graft and right lobe graft; the dimension of surgeon included learning curve within first 90 case and surgeon A; the dimension of post-transplant complications included post-transplantation acute rejection episode, post-transplant hemodialysis (PTHD); the dimension of external environment is team reorganization after 2008-may. The concordance of this statistic model is 0.806.
The risk factors for mortality within 30 days after transplant included: in the dimension of recipient is MELD score > 28; in the dimension of donor-recipient matched is female donate to female; in the dimension of operation is larger blood loss volume; in the dimension of post-transplant complication is the primary non-function of graft (PNF). The area under curve of this statistic model was 0.961. The risk factors of mortality within 90 days after transplant included (1) recipient factor: MELD score > 28 and the serum glucose level > 145 mg/dL while transplant, (2) donor factor: <18 or >26y/o; in the dimension of donor-recipient matched included female donate to female and GRWR<0.9% or >1.8%; (3) operation factor: longer operation time; in the dimension of post-transplant complication are the PNF, the lower FK level in post op 7 day <3.6ng/dl, in hospital complications Clavien grade >= IIIA; (4) surgeon factor: transplant team performed sequence number between 121th and 246th. The area under curve of this statistic model is 0.962.
The risk factors of long term mortality in recipient included (1) recipient-related factor: age <39 or >56y/o, positive history of alcohol drinking, negative history of varices bleeding , MELD score > 28 ,and BMI <20.8 or >32.3 kg/m2; (2) donor-recipient matched factor: female donate to male and GRWR<0.8 % or >1.6% ; in dimension of operation is longer operation time; (3) surgeon factor: learning curve within first 50 case; in the dimension of post-transplant complications are included PTHD, IAI, PNF, abnormal serum sodium level in POD7, DE NOVO cancer, the tuberculosis infection, and acute rejection. The concordance of this statistic model was 0.890.
Conclusion
The risk factors associated with the clinical outcomes of LT are multiple dimensions and multi-factories. After controlling the risk factors of dimensions included donor, recipient, donor-recipient match, and post operation; the learning curve of transplant team & surgeon will affect the overall mortality and post-transplant neurologic complications, biliary complication; but not influent the post transplant early mortality.
Contribution
The present study clarified multiple dimensions that affected the clinic outcomes of LT with ignorance. The results of this study provided information that helps surgeons perform better decision-making, avoid the risk factors and improve the clinical outcomes of LT after surgery. Besides, this study is first study applying R analysis technique to investigate clustered data which was neglected in the past outcome studies in the LT setting. The results of this study encourage the hospital manager to emphasis not only the recipient and donor factors, but also the well-planned training protocol to increment the maturity of transplant team and improve the experience of transplant surgeon.

並列關鍵字

liver transplantation (LT) ； donor-recipient matched factor ； transplant surgeon and team accumulated number ； learning curve ； post-transplant biliary complication ； post-transplant neurologic complication ； early mortality after transplant ； long-term survival ； risk factors ； multivariate regression analysis ； non-linear factor analysis

參考文獻

Abecassis M.M., Burke R., Klintmalm G.B., Matas A.J., Merion R.M., Millman D., Olthoff K., Roberts J.P., American Society of Transplant Surgeons (2009). American Society of Transplant Surgeons transplant center outcomes requirements—a threat to innovation. American Journal of Transplantation, 9(9), 1279-1286.

Akamatsu N., Sugawara Y., Hashimoto D. (2011). Biliary reconstruction, its complications and management of biliary complications after adult liver transplantation: a systematic review of the incidence, risk factors and outcome. Transplantation International, 24(4), 379-392.

Amin M.G., Wolf M.P., TenBrook J.A. Jr., Freeman R.B. Jr., Cheng S.J., Pratt D.S., Wong J.B (2004). Expanded criteria donor grafts for deceased donor liver transplantation under the MELD system: A decision analysis. Liver Transplantation, 10(12), 1468–1475.

Ammori J.B., Sigakis M., Englesbe M.J., O'Reilly M., Pelletier S.J. (2007). Effect of intraoperative hyperglycemia during liver transplantation. Journal of Surgery Research, 140(2), 227-233.

Amodio P., Biancardi A., Montagnese S., Angeli P., Iannizzi P., Cillo U., D’Amico D., Gatta A. (2007). Neurological complications after orthotopic liver transplantation. Digestive and Liver Disease, 39(8), 740–747.