Published online Jan 28, 2017. doi: 10.4254/wjh.v9.i3.147
Peer-review started: September 20, 2016
First decision: October 21, 2016
Revised: November 9, 2016
Accepted: November 27, 2016
Article in press: November 29, 2016
Published online: January 28, 2017
To identify independent risk factors for biliary complications in a center with three decades of experience in liver transplantation.
A total of 1607 consecutive liver transplantations were analyzed in a retrospective study. Detailed subset analysis was performed in 417 patients, which have been transplanted since the introduction of Model of End-Stage Liver Disease (MELD)-based liver allocation. Risk factors for the onset of anastomotic biliary complications were identified with multivariable binary logistic regression analyses. The identified risk factors in regression analyses were compiled into a prognostic model. The applicability was evaluated with receiver operating characteristic curve analyses. Furthermore, Kaplan-Meier analyses with the log rank test were applied where appropriate.
Biliary complications were observed in 227 cases (14.1%). Four hundred and seventeen (26%) transplantations were performed after the introduction of MELD-based donor organ allocation. Since then, 21% (n = 89) of the patients suffered from biliary complications, which are further categorized into anastomotic bile leaks [46% (n = 41)], anastomotic strictures [25% (n = 22)], cholangitis [8% (n = 7)] and non-anastomotic strictures [3% (n = 3)]. The remaining 18% (n = 16) were not further classified. After adjustment for all univariably significant variables, the recipient MELD-score at transplantation (P = 0.006; OR = 1.035; 95%CI: 1.010-1.060), the development of hepatic artery thrombosis post-operatively (P = 0.019; OR = 3.543; 95%CI: 1.233-10.178), as well as the donor creatinine prior to explantation (P = 0.010; OR = 1.003; 95%CI: 1.001-1.006) were revealed as independent risk factors for biliary complications. The compilation of these identified risk factors into a prognostic model was shown to have good prognostic abilities in the investigated cohort with an area under the receiver operating curve of 0.702.
The parallel occurrence of high recipient MELD and impaired donor kidney function should be avoided. Risk is especially increased when post-transplant hepatic artery thrombosis occurs.
Core tip: This retrospective study investigates the occurrence of biliary complications in a total of 1607 consecutive liver transplant patients throughout three decades. Since introduction of Model of End-Stage Liver Disease (MELD)-based liver allocation, the recipient’s MELD-score at transplantation, the development of hepatic artery thrombosis post-operatively, as well as the donor creatinine prior to explantation were identified as independent risk factors, thus a combination of high recipient MELD-score and impaired donor kidney function should be avoided. Risk is especially increased when post-transplant hepatic artery thrombosis occurs. A prognostic model for the prediction of anastomotic biliary complications was developed and successfully internally validated.