Relevance of Preoperative Hyperbilirubinemia in Patients Undergoing Hepatobiliary Resection for Hilar Cholangiocarcinoma

Preoperative hyperbilirubinemia is known to increase the risk of mortality and morbidity in patients undergoing resection for hilar cholangiocarcinoma. The aim of this study was to characterize the associations between the preoperative bilirubin concentration and the risk of postoperative mortality and severe complications to guide decision-making regarding preoperative biliary drainage. Eighty-one patients undergoing liver and bile duct resection for hilar cholangiocarcinoma between 2005 and 2015 were analyzed retrospectively. Postoperative mortality and severe complications, defined as a Clavien–Dindo grade of ≥III, were the primary and secondary outcome measures, respectively. The severe postoperative complications and mortality rates were 28.4% (23/81) and 11.1% (9/81), respectively. Patients with preoperative biliary drainage had significantly lower bilirubin concentrations (p = 0.028) than did those without. The preoperative bilirubin concentration was a risk factor of postoperative mortality (p = 0.003), with an optimal cut-off of 6.20 mg/dL (c-statistic = 0.829). The preoperative bilirubin concentration was a risk factor of severe morbidity (p = 0.018), with an optimal cut-off of 2.48 mg/dL (c-statistic = 0.662). These results indicate that preoperative hyperbilirubinemia is a major risk factor of negative early postoperative outcomes of patients who undergo surgical treatment for hilar cholangiocarcinoma and may aid in decision-making with respect to preoperative biliary drainage.


Introduction
Hilar cholangiocarcinoma is the most prevalent type of malignancy and accounts for 50-60% of all malignancies in the bile ducts [1][2][3]. Extrahepatic bile duct resection, with or without major hepatic resection, with negative surgical margins and potential hepatoduodenal lymphadenectomy remains the gold-standard treatment of patients with hilar cholangiocarcinoma, which has a low response to chemotherapy and radiation therapy [2][3][4]. However, the surgical treatment of hilar cholangiocarcinoma is associated with a relatively high risk of postoperative morbidity and mortality [5]. Preoperative hyperbilirubinemia that is associated with a bile duct stricture appears to be the most important and modifiable risk factor of postoperative mortality and complications [5]. An elevated bilirubin concentration is associated with a higher risk of postoperative liver failure, intraoperative blood loss, postoperative bleeding, infection, heart failure, renal insufficiency, and death in the early postoperative period than a lower bilirubin concentration [5][6][7][8][9][10].
However, the preoperative bilirubin concentration associated with an unacceptable risk of postoperative mortality and severe complications, and the indication for postponing surgery and performing preoperative biliary drainage has not been established [4,5,11]. Preoperative biliary drainage is recommended in patients with malnutrition, hypoalbuminemia, cholangitis, liver failure, or radiochemotherapy, or those who are selected to undergo portal vein embolization [7,12]. This procedure is associated with the risk of iatrogenic cholangitis, postoperative bacteremia, wound infection, longer hospitalization duration, and higher treatment costs [1,2,8,11,13,14].
The association between the preoperative bilirubin concentration and the risk of postoperative mortality and severe complications has not been precisely established yet. The purpose of this study was to precisely characterize the associations between the preoperative serum bilirubin concentration and the risk of postoperative mortality, as well as severe postoperative complications, to propose guidelines for preoperative biliary drainage that would aid decision-making processes.

Materials and Methods
The study protocol was approved by the institutional review board of the Medical University of Warsaw.
In this retrospective, observational study, we evaluated the data of 81 patients who underwent liver and bile duct resection for hilar cholangiocarcinoma at the Department of General, Transplant, and Liver Surgery at the Medical University of Warsaw, Poland, between January 2005 and December 2015. Patients with grade I (Bismuth-Corlette) tumors and those who underwent only extrahepatic bile duct resection were excluded from further investigation.
The primary and secondary outcome measures were postoperative in-hospital mortality and the occurrence of severe postoperative complications that were defined as a Clavien-Dindo grade III or higher, respectively [15]. The preoperative serum bilirubin concentration was the primary factor of interest. Patients were divided into two groups regarding whether or not they underwent preoperative biliary drainage.
The data are presented as numbers with percentages (categorical data) and medians with interquartile ranges (continuous data). Variables were compared using Mann-Whitney's U, χ 2 , and Fisher's exact tests, wherever appropriate, between patients with preoperative biliary drainage and those without, as well as between patients who underwent right or extended right hemihepatectomy and those who underwent left hemihepatectomy. A logistic regression analysis was used to evaluate the risk factors of the established endpoints. Variables were included in multivariable models based on the results of the univariable analyses. The multivariable analysis was conducted using the backward elimination method. The relationship between the preoperative bilirubin concentration and the established endpoints was assessed in subgroups of patients with and without preoperative biliary drainage and patients who underwent right or extended right hemihepatectomy and those with left hemihepatectomy, and subgroups based on the median of other continuous risk factors.
Receiver operating characteristic curves were constructed to discover the optimal cut-off values for the bilirubin concentration to predict the established outcomes in the whole group and those among the subgroups in which the preoperative bilirubin concentration was significantly associated with the analyzed endpoints. Odds ratios (ORs) and areas under the curve (c-statistics) are presented with 95% confidence intervals (95% CIs). The level of statistical significance was set at 0.05. All statistical analyses were performed using Statistica 12 software (Dell Inc., Tulsa, OK, USA).

Discussion
The present study confirms that preoperative hyperbilirubinemia is strongly associated with postoperative mortality and morbidity in patients who undergo surgery for hilar cholangiocarcinoma, especially in those with initially high risk. Based on the obtained results, recommendations for preoperative biliary drainage in patients with hilar cholangiocarcinoma who are qualified to undergo liver and bile duct resection can be made. For clinical use, two cut-off points of the preoperative bilirubin level were established: ≥6.00 mg/dL and ≥2.50 mg/dL. The outcomes of this study indicate that preoperative biliary drainage should be performed in all patients with hilar cholangiocarcinoma with a bilirubin concentration ≥ 6.00 mg/dL, whereas it should be considered in patients with a bilirubin concentration <6.00 mg/dL and ≥2.50 mg/dL, especially in patients with preoperative hypoalbuminemia, anemia, or renal dysfunction, those who are scheduled to undergo right or extended right hemihepatectomy, and those with ineffective biliary drainage. Preoperative biliary drainage is not recommended in patients with a preoperative bilirubin concentration <2.50 mg/dL.
The relationship between preoperative hyperbilirubinemia and the risk of postoperative mortality was strongly confirmed in prior reports. In particular, Jarnagin et al. proved that preoperative hyperbilirubinemia was risk factor of postoperative death in patients undergoing liver resection [16]. Similarly, Belghiti et al. noticed that the mortality rate among patients with preoperative hyperbilirubinemia was significantly higher than that among other patients [17].
In the present study, the optimal bilirubin concentration cut-off to predict postoperative mortality was ≥6.20 mg/dL with a positive predictive value of 29.6% and high sensitivity level of 87.5%, which are already valuable parameters to predict postoperative mortality. The results that were obtained from high-risk patients were even more satisfying. The positive predictive values for the same bilirubin concentration among patients with hypoalbuminemia and anemia and those who underwent preoperative biliary drainage were higher, at 41.2%, 47.1%, and 50.0%, respectively, which indicates that these patients were more likely to die in the early postoperative period than those with lower values. The optimal bilirubin concentration cut-off value to predict postoperative mortality was 2.73 mg/dL in patients who underwent right or extended right hemihepatectomy, with a sensitivity of 100.0% and positive predictive value of 57.1%, which is higher than the value in the whole group, even though the bilirubin concentration was more than twice lower.
Regarding severe postoperative complications, preoperative hyperbilirubinemia was also a significant risk factor, which was widely described in previously published studies [8,18,19]. Miyagawa et al. observed that patients who developed postoperative complications were characterized with a significantly higher preoperative bilirubin concentration than did those without complications [18]. According to Ribero et al.'s research, a preoperative bilirubin concentration >3 mg/dL was an independent risk factor of postoperative liver failure [19], which was a prevalent complication in the studied group. In the present analysis, the optimal bilirubin concentration cut-off to predict severe postoperative complications was ≥2.48 mg/dL in the whole group, with a positive predictive value of 42.5% and sensitivity of 73.9%. The same bilirubin concentration was the optimal cut-off among patients who underwent preoperative biliary drainage and those who required right or extended right hemihepatectomy, with higher positive predictive values of 48.0% and 66.7%, respectively. A higher optimal bilirubin concentration cut-off value of ≥6.20 mg/dL was associated with a higher positive predictive value of approximately 70% in patients with renal dysfunction. Preoperative hyperbilirubinemia led to severe postoperative complications in patients with renal dysfunction, those who underwent ineffective preoperative biliary drainage, and those requiring right or extended right hemihepatectomy. Although the multivariable analysis did not prove that there was a significant association between preoperative hyperbilirubinemia and severe postoperative complications, we do not deny its significant influence, as this could be caused by the limited number of patients in this study and strong influence of renal dysfunction.
The role of preoperative biliary drainage in patients with obstructive jaundice remains controversial [20,21]. However, preoperative biliary tree decompression is recommended in some patients regarding their general condition and comorbidities such as malnutrition, hypoalbuminemia, biliary sepsis, cholangitis, liver insufficiency, and renal failure, as well as in those with preoperative portal vein embolization and chemoradiation therapy [7,12]. Considering the lack of guidelines stating the precise preoperative bilirubin concentration associated with an unacceptably high risk of postoperative mortality and severe complications, as well as the outcomes of the present study, we created the above-named proposal for recommendations for preoperative biliary drainage.
A preoperative bilirubin concentration that indicates the need for preoperative biliary drainage varies considerably among published studies, and this concentration's cut-off has not been analyzed regarding the risk of postoperative mortality and morbidity. Cai et al. concluded that a preoperative bilirubin concentration of 12.40 mg/dL indicated the need for preoperative biliary drainage [22], whereas some researchers claimed that a hyperbilirubinemia level of 10.00 mg/dL required preoperative biliary decompression [5,23]. However, in other reports, a preoperative bilirubin concentration >5.00 mg/dL was the reason for postponing surgery and decreasing the level of hyperbilirubinemia [20]. Additionally, Sano et al. considered a preoperative bilirubin concentration >3.00 mg/dL as an indication for preoperative biliary drainage [24]. Nakanishi et al. assessed an even lower preoperative bilirubin concentration value (>2.00 mg/dL) as an indication for biliary decompression [9].
Hyperbilirubinemia was an additional risk factor of postoperative mortality and morbidity in high-risk patients, and it could be strictly connected to poor outcomes. Hypoalbuminemia, which may coexist with obstructive jaundice because of malnutrition and intestinal absorptive disorders [25], was a significant risk factor of postoperative mortality. Furthermore, hypoalbuminemia is associated with the risk of postoperative liver failure, which may result in postoperative death [19] and was the most frequent reason for death in the analyzed data. In addition, preoperative anemia proved to be a risk factor of postoperative death and morbidity, which was widely observed in other studies [5,[26][27][28]. This comorbidity may result in the need for intraoperative blood transfusions, which were also significantly associated with postoperative mortality in the present study. Similarly, Kennedy et al. found that intraoperative blood transfusions were a risk factor of postoperative liver failure and mortality following liver resection [29]. Moreover, renal insufficiency that was associated with increased preoperative creatinine and urea concentrations enhanced the risk of severe postoperative complications. Squires et al. analyzed data from three hospitals and concluded that patients with a creatinine concentration ≥1.8 mg/dL are at a risk of developing severe postoperative complications, including respiratory insufficiency and renal failure [30]. Finally, patients who undergo right or extended right hemihepatectomy had a higher mortality rate than did those who did not undergo this procedure, and they would probably benefit significantly from preoperative biliary decompression. In other studies, an increase in remnant liver failure because of jaundice was considered the main reason for death in the early postoperative period [7,19]. Other findings prove the necessity of decreasing the level of preoperative hyperbilirubinemia in patients who are qualified to undergo major liver resection [8,11,19,29]. Surprisingly, despite the strong influence of right or extended right hemihepatectomy on the risk of postoperative mortality, the relationship of this surgery with the risk of severe postoperative complications was not observed. The type of liver resection did not influence the occurrence of other severe complications, except for postoperative liver failure, which was the most prevalent reason for postoperative death. Considering the above correlation, patients who are scheduled for right or extended right hemihepatectomy would probably benefit from preoperative biliary drainage. Additionally, although preoperative biliary drainage was not associated with the risk of postoperative mortality and severe complications in the present analysis, the preoperative bilirubin concentration was a risk factor of both outcome measures among patients who underwent preoperative biliary drainage. Therefore, patients with ineffective preoperative biliary drainage would also benefit from decreasing the level of preoperative hyperbilirubinemia because of their probable poor condition and long-lasting hyperbilirubinemia.
In the present study, preoperative biliary drainage was associated with a significantly high positive bile culture rate; therefore, patients who undergo preoperative biliary drainage may carry the risk of contamination. The relationship between preoperative biliary drainage and postoperative infectious complications was described in prior reports [6,21]. Significantly higher gamma-glutamyl transpeptidase activity in patients who underwent preoperative biliary drainage was possibly due to intervention within the biliary tree. Probably, there is no or little relevance of elevated gamma-glutamyl transpeptidase activity among patients with preoperative biliary drainage. However, preoperative bilirubin concentration was a risk factor of severe postoperative complications and mortality in this group of patients, therefore bilirubin concentration should be considered even among patients already with preoperative biliary drainage. Presumably because of selection bias, preoperative biliary drainage was performed more often in patients in worse general condition, mainly in patients requiring right or extended right hemihepatectomy, than in those with a better general condition. Furthermore, patients who underwent preoperative biliary drainage had a bilirubin level that was more than four times lower than that in patients without preoperative biliary drainage, which was indirect proof of the effectiveness of this procedure [6,21]. The presence of a stent in the biliary tree enables the surgeon to identify anatomical structures quickly, resulting in a shorter operation duration in patients who undergo preoperative biliary drainage.
TNM (tumor, nodes, metastasis) stage III or higher did not prove to be a significant risk factor for established endpoints. In the studied group, 14 patients with preoperative biliary drainage and 4 patients without preoperative biliary drainage were characterized with TNM tumor staging III or higher. The analysis included only patients qualified for potentially radical liver and bile duct resection. The small number of patients with advanced TNM stage may result from unresectability of advanced hilar cholangiocarcinoma. In numerous studies, the influence of high TNM stage on early outcomes is not investigated. At the same time, lower AJCC (American Joint Committee on Cancer) T stage has been shown to have positive impact on long-term outcome, including five-year survival [20,31] and disease-free survival [32]. However, long-term outcomes were not the subject of presented study.
Several limitations of the present study must be acknowledged. First, this was a retrospective study; therefore, it was characterized by all of the limitations that are associated with this type of research. Small number of patients included in the analysis may lead to exclusion of other significant relationships. Moreover, the duration of preoperative hyperbilirubinemia, as well as the duration and preferable method of effective preoperative biliary drainage, was not analyzed in the present study. In addition, we did not randomly select patients who were qualified for preoperative biliary drainage, and data about the indications for preoperative biliary drainage were not available; therefore, we did not assess the effectiveness of preoperative biliary drainage. Therefore, a prospective randomized study is warranted.

Conclusions
In conclusion, the outcomes of this study indicate that preoperative hyperbilirubinemia is a major risk factor of poor early postoperative outcomes in patients who undergo surgical treatment of hilar cholangiocarcinoma and may aid in decision-making regarding preoperative biliary drainage. We established a cut-off value to indicate which patients should undergo this procedure. The results indicate that preoperative biliary drainage should be performed in all patients with hilar cholangiocarcinoma with a bilirubin concentration ≥6.00 mg/dL, whereas it should be considered in patients with a bilirubin concentration <6.00 mg/dL and ≥2.50 mg/dL, especially in patients with preoperative hypoalbuminemia, anemia, or renal dysfunction, those who are scheduled to undergo right or extended right hemihepatectomy, and those with ineffective biliary drainage. Preoperative biliary drainage is not recommended in patients with a preoperative bilirubin concentration <2.50 mg/dL.

Conflicts of Interest:
The authors declare no conflict of interest.