Correlation between preoperative biliary drainage, bile duct contamination, and postoperative outcomes for pancreatic surgery

Original communications Correlation between preoperative biliary drainage, bile duct contamination, and postoperative outcomes for pancreatic surgery Paolo Limongelli,a Madhava Pai,a Dev Bansi,b Andrew Thiallinagram,b Paul Tait,c James Jackson,c Nagy A. Habib,a Robin C.N. Williamson,a and Long R. Jiaoa, London, England

Background. Although previously examined, the potential relationship between preoperative biliary drainage (PBD), intraoperative bile culture (IBC), and postoperative morbidity and mortality rate for pancreatic surgery remains unclear. Methods. Two hundred twenty patients underwent operation for either benign pancreatic disease or malignant periampullary and pancreatic neoplasms, consisting of pylorus-preserving proximal pancreatoduodenectomy (n ⫽ 180), biliary bypass (n ⫽ 31), and total pancreatectomy (n ⫽ 9). An intraoperative bile specimen was prospectively collected immediately after division of the bile duct and sent for bacteriologic evaluation for both aerobic and anaerobic microorganisms. Morbidity and mortality rates were evaluated. Results. Of 220 patients evaluated, 113 patients (51.4%) had a positive IBC. Factors associated with a positive IBC were age ⬎70 years (odds ratio [OR], 5.9;95% confidence interval, [CI]: 1.6-22.1; P ⫽ .007), history of coronary artery disease (OR, 0.08; 95% CI, 0.01-0.5; P ⫽ .007), diagnosis of neoplasia (OR, 0.3; 95% CI, 0.1-0.9; P ⫽. 03), and PBD (OR, 0.1; 95% CI, 0.06-0.2; P ⫽ .0001). Infectious complications (OR, 1.8; 95% CI, 1-3; P ⫽ .03), and wound infection (OR, 2.8; 95% CI,1.4-5.3; P ⫽ .002) were greater in patients with positive IBC. Conclusions. PBD predisposes to a positive IBC. Patients with a positive IBC have a clinically important increased risk of developing both infectious complications and wound infection after pancreatic surgery. (Surgery 2007;142:313-8.) From the HPB unit, Hammersmith Campus,a Department of Gastroenterology, Hammersmith Hospital,b and the Department of Radiology,c Hammersmith Hospital, London, England

Biliary tract obstruction is a major factor contributing to the substantial mortality and morbidity of pancreatic surgery.1-4 Under normal circum-

Presented at the 5th Americas Hepato-Pancreato-Biliary Congress, Fort Lauderdale, Florida and at the 7th World Congress of the International Hepato-Pancreato-Biliary Association, Edinburgh, Scotland. Accepted for publication April 15, 2007. Reprint requests: Long R. Jiao, MD, FRCS, Senior Lecturer, Consultant Surgeon, HPB unit, Division of Surgery and Intensive Care, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London, England. E-mail: [email protected] 0039-6060/$ - see front matter © 2007 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2007.04.022

stances, the biliary system is sterile, but in patients with obstructive jaundice, biliary stasis provides fertile ground for bacterial colonization. Biliary instrumentation with or without biliary drainage in patients with obstructive jaundice promotes bile contamination (bacterobilia) and subsequent sepsis.5,6 In fact, virtually all patients with biliary stents develop bacterobilia eventually as a result of ascending colonization, which can lead to postoperative wound infection and septicemia.7,8 This risk is much greater if a stent protrudes into the duodenum.9 Although bacterobilia and septic complications in biliary tract surgery have been examined previously,10 for completeness, the current study was conducted to address the impact of preoperative biliary drainage (PBD) on postoperative complicaSURGERY 313

314 Limongelli et al

tions to examine the potential relationship between PBD, intraoperative bile cultures (IBC), and postoperative morbidity and mortality after pancreatic operation in a tertiary referral center. METHODS Data were collected prospectively on all patients undergoing operation for either benign pancreatic disease or malignant periampullary and pancreatic neoplasms involving 220 patients. Those who had had a biliary bypass procedure before referral to our unit were excluded. Demographic factors and comorbid conditions were recorded, including diabetes mellitus, hypertension, coronary artery disease (CAD), arrhythmia, pulmonary disease, cirrhosis, and alcohol dependence, along with several preoperative laboratory parameters of hematology, biochemistry, and liver function tests. All patients had thorough preoperative staging with a computed tomography or magnetic resonance imaging and postoperative histologic confirmation of diagnosis. PBD was accomplished by either retrograde cholangiopancreatography or percutaneous transhepatic cholangiography. Surgical techniques employed were described previously.11-13 Routinely, before performing the biliary– enteric anastomosis, an intraoperative bile specimen was taken immediately after division of the bile duct and sent for bacteriologic evaluation for both aerobic and anaerobic microorganisms. Bile contamination (bacterobilia) was defined by a positive IBC. To minimize intraperitoneal accumulation of bile, clamping or temporary ligation of the duct was performed during the operation. All patients received prophylactic antibiotics either on induction of anesthesia or preoperatively after stenting (cefuroxime 750 mg and metronidazole 500 mg IV). Biliary instrumentation was covered with antibiotics (ciprofloxacin 400 mg iv). All complications were recorded together with postoperative deaths within 30 days, reoperation in the postoperative period, and duration of hospital stay. After discharge, all patients were followed up at 3-month intervals for the first 6 months, at 12 months after surgery, and then annually. Specifically, the incidence of overall complications, infectious complications, wound infection, intra-abdominal collection or abscess, biliary leak, pancreatic leak, and sepsis syndrome was recorded. Pancreatic fistula was defined as a drain output of any measurable volume of fluid on or after postoperative day 3 with an amylase activity ⬎3 times the serum amylase activity. Statistical analysis. Results are expressed as or mean values ⫾ standard error unless otherwise indicated. Differences in means between subgroups

Surgery September 2007

Table I. Patient’s details, operative procedures, and histologic diagnosis in patients with positive and negative IBC IBC Postoperative parameters Perioperative factors Age (mean ⫾ SD, y) Gender (M/F) Diabetes mellitus (%) Hypertension (%) CAD (%) Arrhythmia (%) Pulmonary disease (%) Cirrhosis (%) Alcohol dependence (%) Obstructive jaundice (bilirubin ⬎20 ␮mol/L) (%) Albumin ⬍30 g/L (%) PBD ERCP (%) PTC (%) Type of procedure PPPD (%) Bypass (%) Total pancreatectomy (%) Histopathologic diagnosis Cancer (%) Chronic pancreatitis (%)

Positive (n ⫽ 113)

Negative (n ⫽ 107)

59.6 ⫾ 1.1 71/42 11.5 13.2 9.7 2.6 4.4 1.7 5.3 58

52.7 ⫾ 1.2* 60/47 14.9 13 1.8* 4.6 2.8 1.8 18.6* 29*

21.2 69 56.6 17.6

14 22.4* 15.8 9.3

50 64.5 33.3

50 35.4 66.6

87.6 15

65.4* 32.7*

IBC, intraoperative bile culture; PBD, preoperative biliary drainage; PPPD, pylorus-preserving proximal pancreatoduodenectomy. *P ⬍ .05 was considered statistically significant.

were compared using Student’s t test or analysis of variance as appropriate. Comparisons between categorical variables were analyzed using the chisquare test. P ⬍ .05 was considered statistically significant. Univariate logistic regression analyses were performed to identify factors associated with bile duct contamination. Univariate factors showing a value of P ⬍ .05 were entered into multivariate logistic regression analysis. Univariate and multivariate associations between either PBD or IBC groups and postoperative outcomes were analyzed via logistic regression analysis. Results are reported as odds ratios (ORs) with 95% confidence intervals (CIs) and probability values. RESULTS Patient characteristics. Patient details, operative procedures, and histologic diagnosis in patients with positive and negative IBC are listed in Table I. Both groups were comparable except for the distribution of age (P ⫽ .0001), CAD (P ⫽ .01), alcohol

Limongelli et al 315

Surgery Volume 142, Number 3

Table II. Main microorganisms isolated from IBC, infected wound, intraabdominal abscess, or blood culture for sepsis No. of positive cultures (%) Microorganisms

IBC

Wound infections

Abscess

Sepsis

ES Lactose-fermenting coliform Non–lactose-fermenting coliform Coliform Streptococcus viridans Yeasts CNS Staphylococcus aureus Pseudomonas species Mixed coliform Enterobacter cloacae Klebsiella Escherichia coli MRSA

47 (40) 42 (36) 11 (9) 10 (9) 10 (9) 8 (7) 8 (7) 7 (6) 6 (5) 6 (5) 5 (4) 3 (3) 3 (3) 3 (3)

9 (16) 15 (14) 4 (3) 6 (5) 0 6 (5) 6 (5) 6 (5) 9 (8) 5 (4) 6 (5) 1 (1) 0 3 (11)

23 (20) 11 (9) 1 (1) 5 (4) 0 3 (3) 6 (5) 2 (2) 6 (5) 0 0 0 0 2 (2)

4 (3) 0 0 0 0 3 (3) 7 (6) 3 (3) 1 (1) 0 2 (2) 2 (2) 0 6 (5)

CNS, coagulase-negative Staphylococcus; ES, enterococcus species; IBC, intraoperative bile culture; MRSA, methicillin-resistant Staphylococcus aureus.

dependence (P ⫽ .002), obstructive jaundice (P ⫽ 0.001), distribution of either cancer (P ⫽ .0001) or chronic pancreatitis (P ⫽ .002), and the number of patients who had PBD (P ⫽ .0001). Patients with PBD had a similar baseline characteristics to those without PBD but a lower rate of alcohol dependence and albumin ⬍30 g/L (P ⫽ .01 and .03, respectively), and a higher rate of cancer and alkaline phosphatase ⬎ 100 U/L (P ⫽ .02 and .0001, respectively). There were no differences in operative time or intraoperative blood loss when patients with positive IBC were compared with those with negative IBC (452 ⫾ 2 minutes vs 460 ⫾ 3 minutes, P ⫽ NS; and 413 ⫾ 16 vs 401 ⫾ 23, P ⫽ NS, respectively). Moreover, the median transfusion rate of red blood cells was also not different when comparing positive (0 U; range, 0-5) with negative (0 U; range, 0-6) IBC (P ⫽ NS) or comparing PBD (0 U; range, 0-4) with no PBD (0 U; range, 0-6; P ⫽ NS). Cultures from patients with infectious complications. Positive IBC were polymicrobial in 46 of 113 patients. The most common isolate was Enterococcus sp (n ⫽ 47) followed by the lactose-fermenting coliforms (n ⫽ 42), non–lactose-fermenting coliforms (n ⫽ 11) and other organisms (Table II). Wound infections were diagnosed by a positive wound swab in 63 patients, of which 26 grew ⬎1 organism. Of these 63 patients, 44 had a positive IBC (69%), and in 36 patients, ⱖ1 of the same microorganism(s) were grown from both bile and wound (81%). Wound infection in patients who had percutaneous transhepatic cholangiography drainage was significantly greater than those who did not (P ⫽ .02).

Table III. Postoperative morbidity in patients with positive and negative IBC IBC

Postoperative parameters

n (%)

Overall complications 101 (45) Infectious complication 90 (41) Wound infection 63 (29) Intraabdominal 42 (19) collection Intraabdominal abscess 35 (16) Biliary leak 8 (4) Pancreatic leak† 19 (11) Sepsis 29 (13)

Positive (n ⫽ 113, %)

Negative (n ⫽ 107, %)

52 49 38 19

39 31* 17* 18

14 3 12 13

17 3 8 14

IBC, intraoperative bile culture. *P ⬍ .05. †Assessed in 180 pancreatoduodenectomies.

Culture-proven abscesses found in 30 patients (13%) were polymicrobial in 17 patients (56%). Of these, 16 had a positive bile culture, and 9 grew the same microorganism(s) from both IBC and the abscess (56%). Septicemia was diagnosed clinically in 29 patients (13%). Of these, 20 had a culture-proven diagnosis (69%) that was polymicrobial in 7 patients, with a positive IBC in 13. Four patients had the same microorganism(s) grown from both the bile and wound cultures. Intraoperative bile culture–related morbidity and mortality. One hundred thirteen patients had a positive IBC (51.4%; Table I). Infectious complica-

316 Limongelli et al

Surgery September 2007

Table IV. Preoperative factors determining the likelihood of positive IBC IBC

Relative risk (95% CI)

Variables

Positive (n ⫽ 113)

Negative (n ⫽ 107)

Univariate

Multivariate

Age (y) ⬎60 ⬎65 ⬎70 CAD Excess alcohol Carcinoma ALP ⬎ 100 Obstructive jaundice PBD

61 45 23 11 6 98 86 61 78

38 21 6 2 20 72 55 33 24

0.4 (0.2-0.8)* 0.3 (0.2-0.6)* 0.2 (0.09-0.59)* 5.6 (1.2-26.1)* 0.2 (0.9-0.6)* 3.1 (1.6-6.2)* 0.3 (0.1-0.5)* 0.3 (0.2-0.6)* 240.1 (0.07-0.2)*

1.3 (0.4-4.3) 1.3 (0.4-4.3) 5.9 (1.6-22.1)* 0.08 (0.01-0.5)* 1.8 (0.5-6.8) 0.3 (0.1-0.9)* 1.8 (0.8-4.1) 0.8 (0.4-1.8) 0.1 (0.06-0.2)*

ALP, alkaline phosphatase; CAD, coronary artery disease; ERCP, endoscopic retrograde cholangiopancreatography; PBD, preoperative biliary drainage. *P ⬍ .05.

tions occurred in 90 patients (40.9%): wound infection (n ⫽ 63), intraabdominal abscess (n ⫽ 35), and septicemia (n ⫽ 29; Table III). Pancreatic leaks developed in 19 of 180 patients after pylorus-preserving proximal pancreatoduodenectomy (PPPD; 10.6%) and bile leaks in 8 (3.6%; Table III). Reoperation was necessary for postoperative complications in 10 patients (4.5%). Among the 210 patients who survived the operation, the median hospital stay was 22 days (range, 9-142) with a morbidity rate of 45.9% (Table III). Postoperative complications occurred in 84 patients after PPPD (46.6%), 12 patients (38.7%) after bypass, and in 5 patients (55.5%) after total pancreatectomy (P ⫽ NS); 10 patients died (4.5%) at a median of 16 (range, 1-40) days postoperative; there were 7 deaths after PPPD (3%), 12 after biliary bypass (6%), and 1 after total pancreatectomy (11%; P ⫽ NS). Multivariate analysis revealed that factors significantly associated with a positive IBC were age ⬎70 years, a history of CAD, a diagnosis of neoplasia, and PBD (Table IV). Seventyeight of 113 (69%) patients with a positive IBC had had a PBD (OR, 0.1; 95% CI, 0.06-0.2; P ⫽ .0001; Table IV), but the rate of both overall and infectious complications were not different in patients with PBD and without PBD (Table IV). A positive IBC did not affect postoperative duration of hospital stay (24 vs 27 days; P ⫽ .11), but the rate of infectious complications and wound infection were significantly greater (Table V). DISCUSSION The current study examined factors contributing to bacterobilia and morbidity and mortality rates in patients undergoing pancreatic surgery for benign and malignant diseases. In most patients who developed postoperative wound infection, the

Table V. Multivariate logistic regression analysis of morbidity and mortality

Overall complications Infectious complication Wound infection Biliary leak Pancreatic leak† Sepsis Mortality

PBD vs no drainage, relative risk (95% CI)

Pos IBC vs Neg IBC, relative risk (95% CI)

1 (0.5-1.7) 1.1 (0.6-2) 1.6 (0.9-3) 1.1 (0.2-4.7) 1 (0.4-2.8) 0.7 (0.3-1.7) 1.1 (0.3-4.1)

1.6 (0.9-2.8) 1.8 (1-3)* 2.8 (1.4-5.3)* 0.9 (0.2-3.8) 1.4 (0.5-3.7) 0.7 (0.3-1.6) 2.2 (0.5-9)

IBC, intraoperative bile culture; PBD, preoperative biliary drainage; NS, not significant. *Independent predictors of outcome by multivariate analysis, P ⬍ .05. †Assesed in 180 pancreatoduodenectomies.

infecting microorganism has been grown from an IBC. Forty percent of cultures were polymicrobial; endogenous biliary pathogens, such as gram-negative aerobes, were most common. Our data are in line with previous studies showing that bacterobilia is an important risk factor for postoperative infectious complications after pancreaticobiliary surgery.14,15 Wells et al6 demonstrated that a preoperative assessment of high-risk factors including age ⬎70 years, the presence of stones, benign stricture in common bile duct, perioperative jaundice, carcinoma, and emergency surgery could not predict accurately the rate of bile colonization. In this study, a strong link was found between several preoperative risk factors and positive IBC. Bacterobilia in patients undergoing biliary tract surgery is one of the factors reported most frequently to be associated with postoperative septic complications. The

Limongelli et al 317

Surgery Volume 142, Number 3

type of microorganisms isolated from the bile was also found identical to those isolated from postoperative infection.16,17 To date, 2 published studies have examined specifically the relationship between positive IBC and postoperative morbidity and mortality in patients undergoing pancreatoduodenectomy. Povoski et al10 reported a series of 161 consecutive patients undergoing pylorus preserving proximal pancreatoduodenectomy over a 3-year period, revealing that 87% of patients with wound infection and 89% with an intraabdominal abscess had a positive IBC. Cortes et al18 compared the outcomes of pancreatoduodenectomy in patients with sterile and infected bile, showing that bile infection was related to previous interventional biliary endoscopy in 80% of patients, which was associated directly with an increased rate of postoperative infections. By using univariate and multivariate models to analyze several end points in the current study, positive IBC significantly increased the risk of postoperative infectious complications and wound infection after pancreatic surgery. Advanced age, obstructive jaundice, carcinoma, and preoperative biliary stenting have all been shown to increase the incidence of bile colonization.19-21 In the present series, the following four independent factors were significantly associated with bile contamination: age ⬎70 years, history of CAD, diagnosis of neoplasia, and PBD. In addition to the reported perioperative risk factors for bile colonization, patients with a history of CAD have also been identified as an important risk factor for bacterobilia.22 The reason for an increased rate of positive IBC in these patients was not examined in this study. Although several studies have addressed the issue of percutaneous and endoscopic insertion of PBD in patients undergoing operation for biliary tract obstruction, the potential advantages and disadvantages continue to be debated.23-25 The current study demonstrated that bacterobilia was most likely to occur in patients with a PBD, which, however, did not influence independently the overall postoperative and infectious complications on multivariate analysis. Sohn et al26 reported a significantly higher incidence of wound infection in patients who had a percutaneous stent insertion before PPPD.26 In the current series, there was a significant association between percutaneous transhepatic cholangiography drainage and wound infection (Figure). At the time of biliary duct transection, colonized bile can lead to contamination of both peritoneal cavity and surgical wound. A previous large prospective cohort study showed no impact on operative morbid-

ity or mortality in patients who had a PBD.27 The effectiveness of PBD in patients with obstructive jaundice secondary to neoplasm was reviewed recently by analyzing previously published, randomized, controlled trials and comparative cohort studies between 1966 and 2001, which demonstrated no benefit in patients with PBD; the authors concluded that PBD should not be performed routinely.28 A randomized, controlled trial might answer questions related to PBD and identify those who would benefit from such intervention preoperatively. But would be a difficult study to carry out because it would require a very large number of patients to detect any meaningful significance.29,30 Limitations. This is a nonrandomized, prospective study with possible selection bias relating to the choice of preoperative biliary decompression and surgeon’s clinical judgment for both the level of bilirubin one would recommend PBD and the timing of operation post PBD. Another drawback is that patients included in this study had 3 distinct surgical procedures that each carried different morbidity and mortality rates. Therefore, it would be more interesting to conduct the study by including a single group of patients, such as those who undergo PPPD. Finally, this is a study from a single tertiary referral center and may not reflect the outcome of PBD in nonreferral hospitals. REFERENCES 1. Armstrong CP, Dixon JM, Taylor TV, Davies GC. Surgical experience of deeply jaundiced patients with bile duct obstruction. Br J Surg 1984;71:234-8. 2. Dixon JM, Armstrong CP, Duffy SW, Davies GC. Factors affecting morbidity and mortality after surgery for obstructive jaundice: a review of 373 patients. Gut 1983;24:845-52. 3. Greig JD, Krukowski ZH, Matheson NA. Surgical mortality and morbidity in one hundred and twenty nine patients with obstructive jaundice. Br J Surg 1988;75:216-9. 4. Pitt HA, Cameron JC, Postier RG, Gadacz TR. Factors affecting mortality in biliary tract surgery. Am J Surg 1981; 141:66-72. 5. Lewis RT, Goodall RG, Marien B, Park M, Lloyd-Smith W, Wigand FM. Biliary bacteria, antibiotic use, and wound infection in surgery of the gallbladder and common bile duct. Arch Surg 1987;122:44-7. 6. Wells GR, Taylor EW, Lindsay G, Morton L. Relationship between bile colonisation, high-risk factors and postoperative sepsis in patients undergoing biliary tract operations while receiving a prophylactic antibiotic. Br J Surg 1989;76:374-7. 7. Yu JL, Andersson R, Ljungh A. Infections associated with biliary drains. Scand J Gastroenterol 1996;31:625-30. 8. Sung JY, Leung JWC, Shaffer EA. Ascending infection of the biliary tract surgical sphincterotomy and biliary stenting. J Gastroenterol Hepatol 1992;7:240-5. 9. Karsten TM, Allema JH, Reinders M, van Gulik TM, De Wit LT, Verbeek PCM, et al. Preoperative biliary drainage, colonisation of bile and postoperative complications in patients

318 Limongelli et al

10.

11.

12.

13.

14.

15. 16. 17.

18.

19.

20.

with tumors of the pancreas head: a retrospective analysis of 241 consecutive patients. Eur J Surg 1996;162:881-8. Povoski SP, Karpeh MS, Conlon KC, et al. Preoperative biliary drainage: impact on intraoperative bile cultures and infectious morbidity and mortality after pancreaticoduodenectomy. J Gastrointest Surg 1999;3:496-505. Williamson RCN. The Pancreas. In: Keen G, Farndon Jr, editors. Operative surgery and management. 3rd ed. Oxford: Butterworth Heinemann; 1994. pp. 161-85. Fleming WR, Williamson RCN. Role of total pancreatectomy in the treatment of patients with end-stage chronic pancreatitis. Br J Surg 1995;82:1409-12. Watanapa P, Williamson RCN. Surgical palliation for pancreatic cancer: developments during the past two decades. Br J Surg 1992;79:8-20. Robson MC, Bogart JN, Heggers JP. An endogenous source for wound infection based on quantitative bacteriology of the biliary tract. Surgery 1970;68:471-6. Chetlin SG, Elliot DW. Biliary bacteremia. Arch Surg 1971;102:303-7. Keighley MRB. Micro-organisms in the bile. Ann R Coll Surg Engl 1977;59:328-34. Pitt HA, Postier RG, Cameron JC. Biliary bacteria. Significance and alterations after antibiotic therapy. Arch Surg 1982;117:445-9. Cortes A, Sauvanet A, Bert F, et al. Effect of bile contamination on immediate outcomes after pancreaticoduodenectomy for tumor. J Am Coll Surg 2006;202:93-9. Keighley MRB, Flinn R, Alexander-Williams J. Multivariate analysis of clinical and operative findings associated with biliary sepsis. Br J Surg 1976;63:528-31. Reiss R, Eliashiv A, Deutsch AA. Septic complications of bile

Surgery September 2007

21.

22.

23.

24.

25.

26.

27.

28.

29. 30.

cultures in 800 consecutive cholecystectomies. World J Surg 1982;6:195-9. Thompson JE, Bennion RS, Doty JE, Muller EIL, Pitt HA. Predictive factors for bacterobilia in acute cholecystitis. Arch Surg 1990;125:261-4. Landau O, Kott I, Deutsh AA, Stelman E, Reiss R. Multifactorial analysis of septic bile and septic complications in biliary surgery. World J Surg 1992;16:962-5. Hatfield AR, Tobias R, Terblanche J, et al. Preoperative external biliary drainage in obstructive jaundice. A prospective controlled clinical trial. Lancet 1982;2:896-9. Lai EC, Mok FP, Fan ST, et al. Preoperative endoscopic drainage for malignant obstructive jaundice. Br J Surg 1994;81:1195-8. Smith RC, Poley M, George CR, et al. Preoperative percutaneous transhepatic internal drainage in obstructive jaundice: a randomised, controlled trial examining renal function. Surgery 1985;97:641-8. Sohn TA, Yeo CJ, Cameron JL, et al. Do preoperative biliary stents increase postpancreaticoduodenectomy complications? J Gastrointest Surg 2000;4:258-68. Bakkevold KE, Kambestad B. Morbidity and mortality after radical and palliative pancreatic cancer surgery. Risk factors influencing the short-term results. Ann Surg 1993;217: 356-68. Sewnath ME, Karsten T, Prins, M, et al. A metanalysis on the efficacy of preoperative biliary drainage for tumour causing obstructive jaundice. Ann Surg 2002;236:17-27. Lillemoe KD. Preoperative biliary drainage and surgical outcome. Ann Surg 1999;2:143-4. Martignoni ME, Wagner M, Krahenbuhl L, et al. Effect of preoperative biliary drainage on surgical outcome after pancreatoduodenectomy. Am J Surg 2001;236:52-9.