Introduction : The present study aims to assess the impact of preoperative biliary drainage on postoperative outcomes and to investigate optimal timing of pancreaticoduodenectomy (PD) following biliary drainage in terms of postoperative morbidity and survival.

Methods : Between January 2007 and December 2015, consecutive 1568 patients underwent PD at a single institute. Clinicopathological and survival data were reviewed retrospectively.

Results : Of all, 831 patients underwent biliary drainage. Mean duration between drainage and surgery was 16.9 days. When the duration was subdivided weekly, 196 underwent PD in 1st week, 300 in 2nd week, 170 in 3rd week, 91 in 4th week, 27 in 5th week and 47 beyond 6th week. In postoperative outcomes, length of stay was longer in drainage group (p=0.028). Postoperative pancreatic fistula was not significantly different between non-drainage and drainage groups (p=0.162), but major complications occurred significantly more in drainage group (p=0.002). Multivariable analysis showed major complications occurred significantly in weekly interval 3 and 4 weeks (odds ratios, 1.863 and 2.523). In survival comparison, there was no difference according to weekly interval in bile duct cancer, but weekly interval of 2 weeks and beyond 6 weeks were associated with poor survival in pancreatic cancer. However, multivariable analysis revealed that the weekly interval was not associated with survival in each cancer.

Conclusions : The interval of 3 or 4 weeks between biliary drainage and PD was associated with major complications, but it was not associated with survival. Therefore, early surgery may enhance postoperative outcomes if preoperative biliary infection is recovered through biliary drainage.