The Impact of Preoperative Biliary Drainage on Postoperative Outcomes after Pancreaticoduodenectomy: A Single Center Retrospective Study

Background The effect of preoperative biliary drainage (PBD) on the short-term outcomes after pancreaticoduodenectomy (PD) remains controversial. Methods 164 consecutive patients with obstructive jaundice who underwent PD in our center from 2016 to 2017 were retrospectively analyzed. The 120 patients who underwent PBD prior to PD (PBD group) were compared with 44 patients who did not (DS group). The short-term outcomes include overall morbidity, severe complications, postoperative pancreatic stula(POPF), postpancreatectomy hemorrhage(PPH), intra-abdominal abscess (IAA), sepsis, delayed gastric emptying (DGE), postoperative hospital stay and 90-day death were assesed. Results There were no signicant statistical differences in overall morbidity, severe complications, POPF, PPH, IAA, LOS and ninety-day mortality between two groups. The incidence of DGE in PBD group was signicantly lower than that in DS group (4.5% vs. 19.2%, P = 0.021). Multivariate regression analysis showed that age over 66 years (OR: 3.094,95% CI: 1.268-7.55) and direct surgery (OR: 5.298, 95% CI: 1.176-23.865) were independent risk factors for DGE. Conclusion For patients with obstructive jaundice, preoperative biliary drainage is independently associated with delayed gastric emptying, but does not affect the overall morbidity and mortality of patients undergoing PD.


Introduction
Periampullary carcinoma includes pancreatic head cancer, ampullary carcinoma, lower common bile duct cancer and duodenal papillary carcinoma. The most symptom is obstructive jaundice, with an incidence of 64-77% [1]. For periampullary carcinoma, the current standard operation is PD. PBD can relieve biliary obstruction, reverse the pathophysiological changes caused by hyperbilirubinemia, such as liver and kidney dysfunction, coagulation abnormality, immune system damage, and so on, thereby affecting the postoperative rehabilitation and long-term survival of patients. Early in 1935, Whipple rst used the method of two-stage operation to treat patients with obstructive jaundice, that is, PD was performed in the second stage after the obstruction was relieved by Cholangiojejunostomy to reduce the postoperative complications and mortality. Henceforth, the concept of PBD has been widely accepted. A study based on SEER database in USA showed that the utilization of preoperative biliary stent placement doubled from 1992 to 2007 [2]. However, since the 21st century, with the continuous progress of surgical technology and perioperative management, more and more studies have shown that PBD could not bene t the patients, on the contrary, it can increase the postoperative complications [3][4][5][6][7][8][9]. But in clinical practice, there are still many situations that patients need PBD due to different reasons. Therefore, it is still controversial whether PBD should be recommended in patients with obstructive jaundice.

Methods
Patients 164 consecutive cases of PD in our center from 2016 to 2017 were retrospectively collected. The patients were divided into PBD group(n = 44) and direct DS group (n = 120) according to whether PBD was performed or not. The demographicscharacteristic perioperative parameters were extracted, including age, gender, comorbidity, operation time, blood loss, pathology, complications, length of hospital stay and ninety-day mortality .

PBD and surgical technique
Preoperative biliary drainage includes percutaneous transhepatic biliary drainage (PTBD) and endoscopic retrograde biliary drainage (ERBD) and endoscopic nasobiliary drainage (ENBD). PD was performed according to the standard procedure, including classic pancreaticoduodenectomy (PD) and pylorus preserving pancreaticoduodenectomy (PPPD). Retroperitoneal Lymphadenectomy was performed in all malignant tumors. In our center, Child method is performed for digestive tract reconstruction and the drainage was placed on the basis of the surgeon's habits.

Statistical analysis
Categorical variables were expressed as the number and percentage and compared using the chi-squared test or the Fisher's exact test as appropriate. Continuous variables were expressed as the mean and standard deviation and compared using the T-test. Univariate and multivariate analyses were performed to nd the risk factors for major morbidity, by using the forward stepwise regression method of maximum likelihood estimation (forward: LR). The parameters which had a statistical signi cance of < 0.1 (indicating a tendency toward signi cance) in univariate analysis were included in a multiple logistic regression model. A two-tailed P-value less than 0.05 was considered statistically signi cant. All statistical analyses were performed by SPSS(version 25.0)

Demographic and Clinical characteristics
A total of 164 consecutive patients were included in the study. The demographic data and clinical characteristics of the two groups are shown in Table 1. There was no statistical difference in average age (PBD 61.8 ± 10.6, DS 61.9 ± 8.8, P = 0.94), gender (108 males, 58 females, P = 0.56), comorbity (hypertension 22.7% in PBD group, 24.2% in DS group, P = 0.848; diabetes 18.2% in PBD group, 16.7% in DS group, P = 0.819). The total bilirubin (TB) before surgery in PBD group was signi cantly lower than that in DS group(PBD 125.9 ± 109.5umol/L,DS 250.6 ± 124.1umol/L P = 0.001) The operation-related data are shown in Table 2. The operation methods were divided into standard PD and PPPD. PD was performed in 84.1% of PBD group, and that of DS group was 85.8%. There was no signi cant difference between the two groups (P = 0.9087). The softness of pancreatic texture is an important factor causing postoperative pancreatic stula. The ratio of soft pancreatic texture of the two groups were 86.4% and 85.8%, respectively, without signi cant difference (P = 0.931). There was no signi cant difference in vascular resection rate between the two groups (PBD 13.6%, DS 26.7%, P = 0.08). In addition, there was no signi cant difference in operation time (PBD 187 ± 68.8 min, DS 187 ± 68.8 min, P = 0.309) and intraoperative blood loss (PBD 485 ± 443 ml, DS 485 ± 443 ml, P = 0.858) between this two groups. The comparison of postoperative complications is shown in Table 3. The incidence of overall morbidity in PBD group was 54.5%, and major complications 20.5%. DS group was 61.7% and 11.7% respectively, in the meanwhile. There was no signi cant difference in the overall morbidity and severe complications between the two groups (P = 0.41, 0.151, respectively). The incidence of POPF was 31.8% in PBD group and 26.7% in DS group, with no signi cant difference (P = 0.515). There was no signi cant difference in PPH (PBD 25%, DS 17.5%, P = 0.283), IAA (PBD 36.4, DS 34.2%, P = 0.793), sepsis (PBD 11.4%, DS 6.7%, P = 0.509), postoperative hospital stay (PBD 19.5 days, DS 18.2 days, P = 0.534) and 90-day death (PBD 6.8%, DS 3.3%, P = 0.588). The incidence of DGE in PBD group was signi cantly lower than that in DS group PBD 4.5%, DS 19.2% (P = 0.021).

Discussion
For patients of resectable periampullary carcinoma with obstructive jaundice, whether to perform biliary drainage prior to PD is still controversial. The previous researches and meta-analysis approached various conclusions. With the development and progress of technology, the drainage technology, operation technology and perioperative management being constantly improved, the necessity of PBD is gradually questioned. In a multicenter retrospective study, 1200 resectable pancreatic cancer patients with obstructive jaundice were included to explore the impact of severe obstructive jaundice on postoperative outcomes, drew a conclusion that serum total bilirubin (TB) > 300 umol / L was an independent risk factor for major morbidity, and was associated with worse prognosis [14]. An RCT study by van der Gaag et al. [3] involved patients with serum total bilirubin between 40-250 umol / L, and it showed that preoperative biliary drainage increased the incidence of complications, suggesting that PBD should not be routinely performed in patients with TB ≤ 250 umol / L. In the study of De Pastena et al. [5], TB > 7.5 mg / dl (128umol / L) without PBD indicated poor postoperative outcomes, and this kind of patients may bene t from PBD. Henceforth, many studies have focused on the necessity of preoperative jaundice.
Most studies have shown that PBD is not really bene cial to patients. Therefore, in the latest guidelines for the diagnosis and treatment of pancreatic cancer (2020 V1.0), which is published by National Comprehensive Cancer Network (NCCN), it is no longer recommended to perform PBD in patients with obstructive jaundice. According to the guidelines, PBD treatment can be considered for patients with following situations: a. cholangitis or fever; b. severe symptomatic jaundice, such as severe pruritus; c.
being delayed for any reason, such as neoadjuvant therapy.
In this study, there was no signi cant difference in the incidence of postoperative overall morbidity and severe complications in the PBD group. Among the various complications, only DGE was related to PBD.
As result shows, PBD can signi cantly reduce the incidence of postoperative DGE. At present, DGE is still a common complication after PD. In patients with obstructive jaundice, the incidence of DGE after PD ranges from 9.4-17.9% [4][5][6]15]. The causes and mechanism of DGE are still unclear. It is still di cult to deal with DGE, without enough effective methods. Based on our results, compared with direct surgery, PBD can reduce the risk of DGEs by 81%, which has a positive effect on promoting the early recovery from the surgery. Through preoperative biliary drainage, preoperative liver function and nutritional status of patients were restored to a certain extent, which may promote the recovery of postoperative gastrointestinal function, thereby reducing the risk of DGE.
There are several limitations in this study. First, this study is a retrospective analysis, confounding factors are unavoidable. Secondly, the sample size of PBD group is small, and the results should be treated with caution. In addition, many patients accepted PBD outside the hospital, and the liver function data before biliary drainage are incomplete, so that the indications of preoperative biliary drainage could not be analyzed in this study. The indication of preoperative biliary drainage is still urgent problem, which needs to be solved in the future.
Owing to the lack of high-quality RCT studies, the necessity of PBD still needs to be explored by prospective study of large sample size.