World Journal of Surgery

, Volume 33, Issue 2, pp 318–325

Preoperative Biliary Drainage Increases Infectious Complications after Hepatectomy for Proximal Bile Duct Tumor Obstruction

Authors

    • Unit of Hepato-biliary-pancreatic and Digestive SurgeryOspedale Mauriziano “Umberto I”
  • Roberto Lo Tesoriere
    • Unit of Hepato-biliary-pancreatic and Digestive SurgeryOspedale Mauriziano “Umberto I”
  • Luca Viganò
    • Unit of Hepato-biliary-pancreatic and Digestive SurgeryOspedale Mauriziano “Umberto I”
  • Luisa Caggiano
    • Unit of Hepato-biliary-pancreatic and Digestive SurgeryOspedale Mauriziano “Umberto I”
  • Enrico Sgotto
    • Unit of Hepato-biliary-pancreatic and Digestive SurgeryOspedale Mauriziano “Umberto I”
  • Lorenzo Capussotti
    • Unit of Hepato-biliary-pancreatic and Digestive SurgeryOspedale Mauriziano “Umberto I”
Article

DOI: 10.1007/s00268-008-9830-3

Cite this article as:
Ferrero, A., Lo Tesoriere, R., Viganò, L. et al. World J Surg (2009) 33: 318. doi:10.1007/s00268-008-9830-3

Abstract

Background

The role of preoperative biliary drainage before liver resection in jaundiced patients remains controversial. The objective of this study is to compare the perioperative outcome of liver resection for carcinoma involving the proximal bile duct in jaundiced patients with and without preoperative biliary drainage.

Methods

Seventy-four consecutive jaundiced patients underwent hepatectomy for carcinoma involving the proximal bile duct from January 1989 to June 2006 and their data were retrospectively analyzed. Fourteen patients underwent biliary drainage before portal vein embolization and were excluded from the study. Thirty patients underwent biliary drainage before hepatectomy and 30 underwent liver resection without preoperative biliary drainage. All patients underwent resection of the extrahepatic bile duct.

Results

Overall mortality and operative morbidity were similar in the two groups (3% vs. 10%, p = 0.612 and 70% vs. 63%, p = 0.583, respectively). The incidence of noninfectious complications was similar in the two groups. There was no difference in hospital stay between the two groups. Patients with preoperative biliary drainage had a significantly higher rate of infectious complications (40% vs. 17%, p = 0.044). At multivariate analysis, preoperative biliary drainage was the only independent risk factor for infectious complication in the postoperative course (RR = 4.411, 95%CI = 1.216-16.002, p = 0.024). Even considering patients with preoperative biliary drainage in whom the bilirubin level went below 5 mg/dl, the risk of infectious complications was higher compared with patients without biliary drainage (47.6% vs. 16.6%, p = 0.017).

Conclusions

Overall mortality and morbidity after liver resection are not improved by preoperative biliary drainage in jaundiced patients. Prehepatectomy biliary drainage increases the incidence of infectious complications.

Introduction

Liver resections in jaundiced patients are associated with high morbidity and mortality rates [1, 2]. Several risk factors have been identified, among which hyperbilirubinemia has been associated with poor early outcome [3, 4]. For these reasons, preoperative biliary drainage (PBD) has been proposed to improve short-term results by reversing pathophysiologic changes observed in jaundiced patients. Several studies tried to demonstrate the benefits of PBD without reaching any clear conclusion; on the other hand, they reported increased morbidity rates in patients with PBD [5]. Moreover, these studies included both benign and malignant bile duct obstructions and analyzed only a small number of hepatic resections.

The aim of this study was to analyze the short-term outcome of patients treated with hepatic and bile duct resection for jaundice due to tumors involving the proximal bile duct, with or without PBD.

Materials and methods

Patients

From January 1989 to June 2006, 74 consecutive patients with jaundice due to tumors involving the proximal bile duct underwent hepatic and bile duct resection in the Unit of Hepato-biliary-pancreatic and Digestive Surgery of Ospedale Mauriziano “Umberto I” of Turin. Clinical data were prospectively collected and retrospectively analyzed. Patients were divided into two groups according to preoperative management: patients who underwent hepatectomy after PBD and those who underwent hepatectomy without drainage. Fourteen patients underwent biliary drainage before portal vein embolization because of a small future remnant liver (FRL) and were excluded from the study. The study focused on 30 patients who underwent PBD before hepatectomy and 30 who underwent hepatic resection without PBD.

Preoperative management

All patients were preoperatively evaluated by a thoracoabdominal computed tomography (CT) scan. Since 1998 all patients with obstructive jaundice have undergone magnetic resonance cholangiography, and those scheduled for major hepatectomy underwent CT volumetry of the future remnant liver (FRL). When the estimated FRL was 30% or less, preoperative portal vein embolization (PVE) was planned and performed after PBD when the serum bilirubin level decreased below 10 mg/dl. Liver resection was performed 4 weeks after PVE if sufficient hypertrophy was obtained.

Indications for PBD were scheduled PVE and signs of cholangitis and malnutrition (serum albumin <3 g/dl). In our center patients underwent PBD by percutaneous transhepatic biliary drainage (PTBD) of the remnant liver, with internal-external drainage whenever possible. Some patients were referred to our center after PBD via PTBD or endoscopic biliary drainage (EBD) at another location. When previous drainage was not adequate, they underwent further drainage by a percutaneous approach.

Surgical procedures

A right subcostal incision extended to the left of the midline was always performed. A midline split was seldom necessary. Since 1997 laparoscopic exploration was performed in patients with hilar cholangiocarcinoma or gallbladder cancer to rule out peritoneal carcinomatosis. Abdominal exploration and intraoperative liver ultrasonography (Aloka SSD1200 with a 7.5-MHz intraoperative linear probe, Aloka CO., LTD, Tokyo, Japan) were always the first step in order to assess resectability. The common bile duct was always divided above the pancreas and a frozen section was always obtained. During liver resection, extrahepatic inflow and outflow vascular control was routinely obtained when possible, with ligation and section of appropriate portal and arterial branches and hepatic veins. Early in the series continuous or intermittent pedicle clamping was performed during parenchymal transection; in recent years no pedicle clamping was performed whenever possible. The parenchymal transection was always performed using a crushing clamp technique and absorbable clips (Absolok® Extra, Ethicon Endo-surgery, Inc., Cincinnati, OH) for any vessel or bile duct. Since 1999 we have used bipolar forceps with continuous irrigation before section of small vessels and the number of absorbable clips has decreased. In case of hilar cholangiocarcinoma, resection of segment 1 was systematically performed. In cases of gallbladder cancer or hilar cholangiocarcionoma, lymphadenectomy of the first and second levels was always performed. Biliary reconstruction was performed by Roux-en-Y hepaticojejunostomy or intrahepatic cholangiojejunostomy. Systematic frozen-sections samples of the proximal bile ducts were obtained from patients with hilar cholangiocarcionoma or gallbladder cancer. One or two abdominal drains close to the raw cut surface and the biliary anastomosis were usually used.

Definitions

Jaundice was defined as total serum bilirubin level higher than 2 mg/dl. Major hepatectomy was defined as resection of three or more Couinaud segments. Extended hepatectomy was defined as resection of five or more Couinaud segments. Types of hepatectomies were classified according to Brisbane 2000 terminology [6]. Preoperative biliary drainage were defined as EBD, PTBD, or surgical drainage. Operative mortality was defined as death within 60 days after surgery or before discharge from hospital.

Morbidity included any deviation from the normal postoperative course. Major morbidity was defined as a grade III or higher complication according to the classification scheme proposed by Dindo et al. [7]. Complications were classified as infectious or noninfectious in nature. Infectious complications recorded were wound infections (superficial or deep infections that required antibiotics and/or wound drainage), intra-abdominal abscess (intra-abdominal collection associated with fever and/or leukocytosis that required drainage, culture of which yielded positive findings), lung infections (infiltrate on chest radiogram requiring antibiotic treatment), cholangitis (positive finding of a bile culture associated with fever), and sepsis (positive findings of hemoculture associated with fever and/or leukocytosis). Liver failure was defined as PT <50% after postoperative day 4 and a lack in reduction of serum bilirubin. A bile leak was defined as the drainage of 50 ml or more of bile from the surgical drain or from drainage of an abdominal collection, over a period of 3 days or more [8].

Statistical analysis

Continuous variables were compared between groups by the unpaired t test or the Mann–Whitney U test as appropriate; categorical variables were compared by the χ2 test or Fisher’s exact test as appropriate. A p value of less than 0.05 was considered significant for all tests. Multivariate analysis was performed by analyzing all the variables significant at univariate analysis (p < 0.05) in a multiple regression.

Results

Preoperative characteristics

Preoperative characteristics are given in Table 1. Mean age at time of surgery was 64.8 years (range = 35-80). The most common cause of biliary obstruction was hilar cholangiocarcinoma in 44 patients (73%); among these, two patients had a final histologic diagnosis of benign stricture. Causes and level of biliary obstruction were similar in both groups. Liver function tests at initial presentation were similar in both groups. Median serum bilirubin level was 12.3 mg/dl (range = 3.5-31.2) in the PBD group and 13.2 mg/dl (range = 2.1-31.8) in the no-drain group (p = 0.499). Preoperative serum bilirubin level was significantly lower in the PBD group after drainage (median = 3.1 vs. 13.2 mg/dl, p < 0.0001).
Table 1

Preoperative characteristics

 

PBD (30 pts)

No PBD (30 pts)

p

Patients characteristics

   Age (mean, range)

66.0 (47–80)

63.5 (35–80)

0.318

   Sex (M/F)

13/17

19/11

0.120

   Bilirubin (mg/dl, median)

      Before drainage

12.3 (3.5–31.2)

13.2 (2.1–31.8)

0.499

      After drainage

3.1 (0.3–14.1)

<0.0001

Cause of biliary obstruction

 

   Hilar cholangiocarcinoma

26 (86.7%)

18 (60%)

0.410

   Gallbladder cancer

3 (10%)

6 (20%)

0.469

   Intrahepatic cholangiocarcinoma

1 (3.3%)

2 (6.7%)

1

   Hepatocellular carcinoma

0 (0%)

2 (6.7%)

0.491

   Colorectal metastasis

0 (0%)

1 (3.3%)

1

   Biliary papillomatosis

0 (0%)

1 (3.3%)

1

Preoperative biliary drainage

Twenty-six patients underwent PBD before referral to our center. Four patients underwent PTBD in our center because of poor general condition, namely, malnutrition with serum albumin lower than 3 g/dl.

In the PBD group, ten patients (33%) were treated with one or more EBD before referral. Among these patients, three underwent further PBD by PTBD because of inadequate decrease of serum bilirubin (30% of patients with EBD). Twenty-one patients (70%) were treated by PTBD: 14 (67%) by internal-external drainage and 7 (33%) by external drainage. Two patients underwent biliary drainage at the time of abdominal exploration performed elsewhere: one with a T-tube and one with two stents placed through a duodenotomy.

Median duration of PBD (meant as the first procedure in those patients who required further drainage) was 27.5 days (range = 10–90). In 21 patients (70%) serum bilirubin level decreased below 5 mg/dl. Drainage-related morbidity was 23% (7 patients). No major morbidity was recorded. Six patients (20%) developed an infectious complication (cholangitis): one patient with PTBD, one patient with surgical drainage, and four with EBD (three required further PTBD).

Surgical procedures

Type and extent of liver resection and number of resected liver segments were similar in the two groups (Table 2). All patients underwent resection of the biliary confluence, except two patients, both without PBD, who underwent bisegmentectomy Sg4b-5 and main bile duct resection for gallbladder carcinoma.
Table 2

Surgical procedures

 

PBD (30 pts)

No PBD (30 pts)

p

Major hepatectomy

28 (93%)

25 (83%)

0.423

Extended hepatectomy

13 (43%)

10 (33%)

0.425

Right or extended right hepatectomy

14 (47%)

8 (27%)

0.108

Left or extended left hepatectomy

13 (43%)

17 (57%)

0.301

Resected segments (median)

4 (2–6)

4 (2–6)

0.478

Pedicle clamping

10 (33%)

11 (37%)

0.786

Biliary convergence resection

30 (100%)

28 (93%)

0.491

Reconstructed biliary ducts

   1 duct

14 (47%)

8 (27%)

0.108

   ≥2 ducts

16 (53%)

22 (73%)

Associated resections

4 (13%)

6 (20%)

0.729

Vascular resection

4 (13%)

9 (30%)

0.210

Radical resection

28 (93%)

28 (93%)

1

Biliary procedures included Roux-en-Y hepaticojejunostomy (2 patients) or intrahepatic cholangiojejunostomy with reconstruction of one to five bile ducts (58 patients); no difference in surgical procedures was recorded in the two groups. A similar number of patients underwent vascular resection in both groups: four patients in the PBD group underwent portal vein resection (one case associated with hepatic artery resection and reconstruction), four patients in the no-drain group underwent portal vein resection, and two patients underwent hepatic artery resection and reconstruction.

In four PBD patients hepatectomy was associated with other resections: one pancreaticoduodenectomy, one left adrenalectomy, and one gastric and one duodenal partial resection. In the no-drain group, one patient underwent pancreaticoduodenectomy and one patient had a peritoneal node that was removed. Pedicle clamping was similarly carried out in the two groups.

Short-term outcomes

Mortality was 7%, without significant differences between the two groups (Table 3). One patient in the PBD group died because of sepsis and liver failure. In the no-drain group three patients died of liver failure. Morbidity (70% vs. 63%, p = 0.583) and major complication rates (37% vs. 53%, p = 0.194) were similar in the two groups. A significantly higher number of patients in the PBD group developed infectious complications (40% vs. 17%, p = 0.044). No difference in infectious complication rate was recorded between patients with EBD and PTBD (57% vs. 64%, p = 0.729). When drainage procedure-related complications were included, infectious morbidity increased further in the PBD group (53% vs. 17%, p = 0.001). Noninfectious morbidity rates were similar in the two groups (56% vs. 53%, p = 0.795).
Table 3

Postoperative outcomes

 

PBD (30 pts)

No PBD (30 pts)

p

Mortality

1 (3%)

3 (10%)

0.612

Morbidity

21 (70%)

19 (63%)

0.583

Major morbidity

11 (37%)

16 (53%)

0.194

Infectious morbidity

12 (40%)

5 (17%)

0.044

Number of infectious complications

17

9

0.193

   Wound infection

1 (3%)

1 (3%)

1

   Intra-abdominal abscess

4 (13%)

2 (7%)

0.670

   Lung infection

6 (20%)

3 (10%)

0.469

   Cholangitis

3 (10%)

1 (3%)

0.612

   Sepsis

3 (10%)

2 (7%)

1

Noninfectious morbidity

17 (56%)

16 (53%)

0.795

Number of noninfectious complications

19

20

0.865

   Bleeding

3 (10%)

2 (7%)

1

   Collection

4 (13%)

4 (13%)

0.704

   Bile leak

7 (23%)

10 (33%)

0.390

   Liver failure

5 (17%)

4 (13%)

1

Overall morbidity

24 (80%)

19 (63%)

0.152

Overall major morbidity

11 (37%)

16 (53%)

0.194

Overall infectious morbidity

16 (53%)

5 (17%)

0.002

Blood transfusions

21 (70%)

17 (57%)

0.283

Reoperation

4 (13%)

3 (10%)

1

Hospital stay

28.1 (9–85)

29.2 (7–120)

0.855

Postoperative liver function tests are depicted in Fig. 1. Prothrombin activity decreased on postoperative day 1 in both groups (significantly lower in the PBD group, p = 0.022) and gradually increased on the following days. Postoperative changes in AST and ALT levels were similar in both groups. Serum bilirubin level was higher in the preoperative assessment in the no-drain group and lost significance from postoperative day 10. Blood transfusion rates were similar in the two groups. There was no significant difference in postoperative hospital stay.
https://static-content.springer.com/image/art%3A10.1007%2Fs00268-008-9830-3/MediaObjects/268_2008_9830_Fig1_HTML.gif
Fig. 1

Postoperative liver function tests. Prothrombin activity (PT) was significantly lower in PBD group (p = 0.022) in postoperative day 1, and was similar the following days. AST and ALT levels were similar in both groups. Serum bilirubin level was significantly higher in the no drain group (postoperative day 1: p = 0.00005; postoperative day 3: p = 0.0001; postoperative day 7: p = 0.01). Difference lost significance from postoperative day 10 (p = 0.216)

To evaluate the role of efficacy of PBD, patients in the PBD group were divided in two subsets: patients with preoperative bilirubin lower (21 patients) and higher (9 patients) than 5 mg/dl. No differences were recorded in mortality and major or infectious morbidity. Comparing PBD patients with bilirubin lower than 5 mg/dl with no-drain patients (Table 4), significantly higher infectious morbidity rates were observed in the PBD group (48% vs. 17%, p = 0.017).
Table 4

Comparison of PBD group patients with bilirubin <5 mg/dl vs. no-PBD group patients

 

PBD bilirubin <5 mg/dl (21 pts)

No PBD (30 pts)

p

Mortality

1 (5%)

3 (10%)

0.634

Morbidity

14 (67%)

19 (63%)

0.806

Major morbidity

9 (43%)

16 (53%)

0.461

Infectious morbidity

10 (48%)

5 (17%)

0.017

Noninfectious morbidity

11 (37%)

16 (53%)

0.946

Risk factors associated with infectious complications at univariate analysis were PBD (p = 0.045), preoperative bilirubin lower than 5 mg/dl (p = 0.053), and preoperative ALT higher than 100 U/l (p = 0.035). Multivariate analysis identified PBD as the only independent risk factor for infectious complications (RR [relative risk] = 4.411, 95%CI [confidence interval] = 1.216–16.002, p = 0.024) (Table 5).
Table 5

Risk factors for infectious complications

 

No. of patients

Incidence of infectious complications

Univariate

Multivariate

p

Odds ratio

p

Age (years)

   <65

26

10 (39%)

0.127

  

   >65

34

7 (21%)

  

Sex

   M

32

8 (25%)

0.540

  

   F

28

9 (32%)

  

PBD

   Yes

30

12 (40%)

0.045

4.411 (1.216–16.002)

0.024

   No

30

5 (17%)

1

Preoperative bilirubin <5 mg/dl

   Yes

27

11 (41%)

0.053

1.941 (0.464–8.021)

0.364

   No

33

6 (18%)

1

AST > 100 U/l

   Yes

17

3 (18%)

0.402

  

   No

43

14 (33%)

  

ALT > 100 U/l

   Yes

28

4 (14%)

0.049

0.346 (0.091–1.314)

0.119

   No

32

13 (41%)

1

Major hepatectomy

   Yes

53

15 (28%)

1

  

   No

7

2 (29%)

  

Extended hepatectomy

   Yes

23

5 (22%)

0.371

  

   No

37

12 (32%)

  

Pedicle clamping

   Yes

21

6 (29%)

0.976

  

   No

39

11 (28%)

  

Vascular resections

   Yes

13

3 (23%)

0.898

  

   No

47

14 (30%)

  

Associated resections

   Yes

6

3 (50%)

0.445

  

   No

54

14 (26%)

  

No. of bile duct reconstructed

   1

22

4 (18%)

0.303

  

   ≥2

38

13 (34%)

  

Blood transfusion

   Yes

38

13 (34%)

0.303

  

   No

22

4 (18%)

  

Discussion

Liver surgery in jaundiced patients is supposed to carry particular risks because of hepatic and systemic changes caused by hyperbilirubinemia [1, 2]. Experimental studies on jaundiced animals showed the benefits of biliary drainage, especially internal biliary drainage with restoration of enterohepatic circulation of biliary salts [9, 10]. The aim of PBD should be to increase cholestatic liver tolerance to ischemia [11] and to reduce blood loss [12]. Several studies have been carried out to demonstrate the benefits of PBD but did not show any clear evidence of advantages [1316]. Most of these studies lacked uniformity, analyzing both lower and upper biliary obstructions and benign and malignant lesions, and they included only a small number of hepatic resections. Nevertheless, many authors advocate the systematic use of PBD before hepatic resection [1719], reporting low mortality and morbidity rates.

Cherqui et al. [20], in a case-control study comparing jaundiced patients treated with major hepatic resections for carcinoma of the proximal biliary duct with nonjaundiced patients, showed in the former group higher rates of complications (mainly bile leaks and subphrenic collections) and higher transfusion rates, even though the rates were comparable to those reported in the literature in patients with PBD. Nevertheless, 25 of 27 control cases were hepatectomies without bile duct resection, which usually have lower morbidity rates. In our study, all procedures were hepatic and bile duct resections. Furthermore, most of the resections were major hepatectomies. Analysis of the short-term results showed similar morbidity and mortality rates (including bile leak, collections, and liver failure). Moreover, it should be remembered that drainage procedures are not risk-free, especially for proximal bile duct obstructions: morbidity rates of 25–50% and mortality rates of 3–5% [21, 22] for endoscopic drainage and of 5–20% and 0–5% for percutaneous transhepatic drainage are reported [2325]. In our study, drainage-related morbidity was 23%. It should be stressed that only 4 of 30 patients in the PBD group underwent drainage following our indication. Ten patients received EBD which was inadequate in three (30% of patients with EBD), and four developed cholangitis. EBD should not be recommended as the procedure of choice for PBD as long as it is associated with high complication rates for proximal bile duct obstruction and short patency time [21] that may require stent removal and replacement. Besides, PTBD also has risks, even in skilled hands. Ishizawa et al. [26] reported a series of 42 patients treated with selective or total PBD in which the overall preoperative cholangitis rate was 31% and the postoperative cholangitis rate was 26%.

Bacterial contamination following biliary drainage is indeed a crucial feature. Hochwald [27] showed that biliary stents (more likely endoscopic stents rather than percutaneous stents) are associated with bacterobilia and that patients with PBD develop infectious complications significantly more often than patients without PBD. That study analyzed data from patients with or without PBD who were treated with both curative surgery and palliative biliary bypass; among the curative operations, only 22 were hepatic resections. Our study analyzed exclusively patients with hepatic resection and showed that a higher number of patients with PBD developed infectious complications compared with the no-drain group.

PBD was performed by different techniques (PTBD, EBD, or surgical BD) because most of the patients were referred to our center after PBD had been performed. EBD is commonly associated with higher rates of bacterial contamination [27, 28]. Nevertheless, in our series EBD patients had an infectious complications rate similar to that of PTBD patients, confirming that the higher number of postoperative infections in PBD patients is not related to the drainage technique but to the drainage itself.

Comparing PBD patients who had a preoperative serum bilirubin level below 5 mg/dl with no-drain patients, patients with preoperative drainage experienced a higher number of infectious complications. This analysis showed that infectious morbidity is not related to the incomplete resolution of jaundice but to the presence of the biliary drainage, which was the only independent prognostic factor for the onset of postoperative infectious complications in the multivariate analysis. When cholangitis following drainage procedures is included, the rate of infectious morbidity becomes even higher.

In summary, mortality and morbidity rates after liver resection are not improved by preoperative biliary drainage in jaundiced patients. Prehepatectomy biliary drainage increases the incidence of infectious complications, even when a preoperative bilirubin level less than 5 mg/dl is achieved.

Copyright information

© Société Internationale de Chirurgie 2008