A Matched Case-Control Study of Preoperative Biliary Drainage (Pdb) in Patients with Pancreatic Adenocarcinoma: Routine Pbd Is Not Justified
James J. Mezhir*1, Murray F. Brennan1, Raymond Baser2, Michael D'Angelica1, Yuman Fong1, Ronald P. Dematteo1, William R. Jarnagin1, Peter J. Allen1
1Surgery, Memorial Sloan Kettering Cancer Center, New York, NY; 2Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY
Introduction: Previous studies of PBD in patients with pancreatic malignancy and jaundice suggest an increase in perioperative complications. Due to the conflicting results of prior studies, PBD remains controversial.Methods: Patients who underwent pancreaticoduodenectomy (n=340) without neoadjuvant therapy or prior biliary bypass for pancreatic adenocarcinoma between 2000 and 2005 were included. Perioperative complication data were obtained from a prospectively maintained database. To perform a matched case-control analysis, 94 PBD and 94 non-stented (no PBD) patients were matched for age, gender, preoperative albumin, and bilirubin levels (PBD group: pre-stent bilirubin; no PBD group: preoperative bilirubin). Results: Overall, PBD was performed in 201 patients (60%), 178 patients (89%) had internal drainage and the remaining internal/external stents. Indications for stenting were obstructive jaundice (96%) and cholangitis (3%). Stent-related complications including pancreatitis, duodenal perforation, and cholangitis or blockage requiring stent revision occurred in 47 patients (23% of PBD group). Median time from PBD to resection was 27 days (range 4-157). After matching, there was no statistically significant difference in operative time, perioperative blood loss or transfusions, postoperative INR, or hospital stay. Intraoperative bile cultures were positive in 83% of PBD patients vs. 7% no PBD (p=0.001). The postoperative morbidity rate was 44.1% and the mortality rate, 3%. There was a significant increase in infectious complications in the PBD group (23.4%) vs. no PBD (9.6%, p=0.009). Intraabdominal abscess occurred in 12% of PBD patients compared to 2% of no PBD (p=0.01). Wound infections occurred in 21% of the PBD group vs. 7% of no PBD (P=0.007). There was no difference in the number of anastomotic leaks or fistulas (p=0.53). Conclusions: PBD does not improve outcome, is associated with a stent-related complication rate of 23%, and resulted in a two-fold increase in post-pancreatectomy infectious complications. The routine use of PBD is not justified.
Variables in matched patients
| No PBD (N=94) | PBD (N=94) | P-value | |
| Median pre-stent bilirubin (range) | - | 10.9 (0.6-31.2) | - |
| Median preoperative bilirubin (range) | 11.4 (0.2-36.6) | 1.9 (0.4-27.4) | 0.001 |
| Median preoperative albumin (range) | 4.0 (2.5-4.8) | 3.9 (1.3-4.9) | 0.510 |
| Positive intraoperative bile culture | 7% | 83% | 0.001 |
| Overall complication rate | 41% | 45% | 0.440 |
| Anastomotic leak or fistula | 6% | 4% | 0.530 |
| Infectious complications | 10% | 23% | 0.009 |
| Intraabdominal abscess | 2% | 12% | 0.010 |
| Wound infection | 7% | 21% | 0.007 |
Back to Program | 2009 Program and Abstracts | 2009 Posters
