Goal-oriented preoperative biliary drainage is more precise and conducive to seize the opportunity for pancreaticoduodenectomy

Abstract

Background

Preoperative biliary drainage (PBD) for selected patients with severe juandice has been shown to improve clinical conditions for pancreaticoduodenectomy (PD) and reduce the risk of post-pancreatectomy hemorrhage (PPH). However, the determination of an optimal end-point for PBD remains unclear. The aim of this research is to introduce the concept of goal-oriented biliary drainage, which may serve as a reasonable target and identify the optimal surgery time window.

Methods

The clinical data of 194 patients diagnosed with pancreatic cancer and obstructive jaundice were retrospectively analyzed. Serological laboratory examinations including total bilirubin (TBIL) within one week before PBD and PD were recorded and labeled as TBIL-pre and TBIL-post respectively. PBD and PD were performed by experienced medical teams. PPH with grade B and C were enrolled.

Results

TBIL-post less than 93.0µmol/L (sensitivity 83.78%, specificity 72.61%) or TBIL decay more than 68.5% (sensitivity 86.49%, specificity 69.43%) identified through ROC curves and multivariate analysis were independent protective factors for reducing the risk of PPH (OR 0.234 and 0.191 retrospectively, P<0.05) and were established as PBD goals. The Kaplan-Meier curves demonstrated the median time to achieve both PBD goals was 3 weeks. Additionally, the proportion of patients achieving both goals failed to increase with the PBD duration over 6 weeks (P > 0.05). The proportion of TBIL-post ≤ 93.0µmol/L (70.8% vs. 51.1%, P<0.05) and TBIL decay ≥ 68.5% (67.0% vs. 50.0%, P<0.05) were higher in EBS group than those in PTCD group.

Conclusion

A goal-oriented PBD with the target of TBIL ≤ 93.0µmol/L or TBIL decay ≥ 68.5% can reduce the morbidity of PPH. In general conditions, PBD duration within 3 weeks would be sufficient, while exceeding the duration beyond 6 weeks could not provide additional benefits. Both EBS and PTCD are safe and EBS is more recommended due to its superior performance in achieving the goals.

Introduction

Pancreaticoduodenectomy (PD), the most effective curable treatment for carcinoma of the pancreatic head or uncinate process, unfortunately is still associated with a high postoperative morbidity ranging from 35 to 45%, despite of a significant reduction in mortality to below 5% due to advancements in surgical techniques [1]. Obstructive jaundice secondary to pancreatic head carcinoma possibly causes biliary infection or hemorrhage attributing to cholestasis, hepatocyte damage or coagulation function deficits.

According to NCCN guidelines, preoperative biliary drainage (PBD) may be considered for patients with symptoms of cholangitis/fever or severe symptomatic jaundice (intense pruritus), or when surgery is delayed for any reason, such as the administration of neoadjuvant therapy, although PBD is not routinely recommended prior to planned surgery [2]. The improvement of perioperative safety and adaptability for surgery is supposed to be the dominant purpose of PBD, and a prolonged PBD aimed at excessively lowering or normalizing bilirubin levels should be avoided so as not to miss the optimal window for surgery, considering the rapid progression of pancreatic cancer.

Concerning the timing of PBD, numerous previous studies have addressed the question of “when to start”, leaving the confusion of “when to stop” to be settled. In this research, the concept of the goal-oriented biliary drainage is introduced, which aims to provide guidance for setting reasonable objectives for PBD and identifying the optimal window for surgery.

Methods and materials

Patient selection

According to the NCCN guidelines and the clinical practices, PBD was performed in patients with symptoms of severe cholangitis, intense pruritus, a delayed surgery or other necessary conditions. The medical records of patients pathologically diagnosed with pancreatic cancer and obstructive jaundice who underwent PBD and PD from 2016 to 2023 in Peking Union Medical College Hospital were retrospectively reviewed. Exclusion criteria: (1) Mild obstructive jaundice, defined as Total bilirubin (TBIL) < 171µmol/L before PBD; (2) Receipt of neoadjuvant therapy; (3) Only biopsy or palliative surgery; (4) Patients who underwent either PBD or PD at other centers and subsequently transferred to our hospital.

Methods and research parameters

Serological laboratory examinations within one week before PBD were recorded including TBIL, Albumin (ALB), Alanine aminotransferase (ALT), Hemoglobin (HGB), White blood cell count (WBC), Percentage of neutrophil (NEUT%), Platelet count (PLT), Prothrombin time (PT), Activated partial thromboplastin time (APTT), International normalized ratio (INR), and were labeled as ‘examination-pre’, such as TBIL-pre. The selection of PBD, including endoscopic biliary stenting (EBS) or percutaneous transhepatic cholangial drainage (PTCD), was based on the Tumor anatomical features and determined by the clinical judgment from a Multidisciplinary Consultation and specific patient circumstances at the time of treatment. EBS was typically preferred in cases when patients were in need of biopsy. When the attempt of EBS failed, PTCD would be considered as a remedial measure. Serological laboratory examinations mentioned above were dynamically monitored after PBD, and those conducted within one week prior to pancreaticoduodenectomy were labeled as ‘examination-post’, such as TBIL-post. The decisions regarding the surgical approach of open or laparoscopic were based on the clinical judgment from the chief surgeon. Open surgery would be indicated in more complex cases, such as those involving extensive disease, concerns about patient safety or laparoscopic techniques. Both open and laparoscopic surgeries were performed by experienced surgery teams, by whom equivalent surgical quality was guaranteed (Supplementary Table 1) by chief surgeons with experience in over 30 cases who have surpassed their learning curves on the basis of informed consent from patients. All selections were conducted with informed consent from the patients.

Definition criterion

Postoperative morbidity was recorded within a time interval of 30 days after surgery, including postoperative pancreatic fistula (POPF) and post-pancreatectomy hemorrhage (PPH), and was defined according to the criteria of International Study Group of Pancreatic Surgery [3,4,5]. In this research, clinically relevant POPF (CR-POPF, POPF grade B and C) was inclusive into postoperative morbidity, and similarly only PPH grade B or C were inclusive. Surgical site infection (SSI) was defined as an infection that occurred within 30 days postoperatively affecting the skin or subcutaneous tissue of the incisional region or deeper soft tissues according to Guidelines for Prevention of SSI published by the Centers for Disease Control [6].

In this research, abnormal biliary drainage events included PTCD tube blockage or detachment, biliary stent occlusion or migration, leading to the need for further invasive treatment. PBD relevant morbidity included biliary infection, hemorrhage, ERCP relevant pancreatitis and abnormal biliary drainage events.

Statistical analysis

Statistical analyses were performed using SPSS 25.0 (SPSS Inc., Chicago, IL, USA). Continuous variables were analyzed by Mann-Whitney U test and were presented as median (Inter quartile range, IQR), while categorical variables were analyzed by Chi-square or Fisher exact test as appropriate, and were reported as frequency (percentages). Predictive factors for postoperative morbidity were identified using a multivariable logistic regression model (inclusive criterion: 1.univariate P-value<0.2; 2. accordance of clinical significance). The incidence of achieving PBD goals was estimated using the Kaplan-Meier method, and differences were evaluated with log-rank test. Statistical significance was defined as a P-value of less than 0.05.

Results

Demographic, PBD, surgery and morbidity data

A total of eligible 194 patients were included in this study (Table 1). The median age of the patients was 63 (56–68) years. 105 (54.1%) patients were pathologically diagnosed with stage II pancreatic carcinoma, followed by 48 (24.8%) patients with stage III and 41 (21.1%) patients with stage I. The median PBD duration was 15 (10–27) days. Among the patients 106 (54.6%) underwent EBS while 88 (45.4%) received PTCD. 151 (77.8%) cases of laparoscopic pancreaticoduodenectomy were performed while 43 (22.2%) direct open or conversion to open surgeries were implemented. PPH were observed in 37 (19.1%) cases, including 20 (54.1%) grade B and 17 (45.9%) grade C. In addition, 37 (19.1%) cases of CR-POPF were identified, with 33 (89.2%) cases classified as grade B and 4 (10.8%) as grade C. A total of 73 (37.6%) patients developed SSI, among which 64 (87.7%) cases were intra-abdominal infection, 14 (19.2%) were incision infection, and 5 (6.9%) involved complex infection of both intra-abdomen and incision.

Table 1 Demographic, PBD, surgery and morbidity data

Exploration of TBIL Levels as a target for reducing the PPH risk

To explore PBD and TBIL-related factors affecting PPH, 194 patients were divided into PPH group (37 individuals) and Non-PPH group (157 individuals). Univariate analysis of clinical characteristics revealed significant differences in CR-POPF, TBIL-post, TBIL decay and PT-post (P<0.05) (Table 2). In addition to these factors, PBD methods (EBS/PTCD), PBD duration, ALT-post and APTT-post (P<0.2) were also considered potential factors for multivariate analysis. No significant differences were found in other factors or baseline serological results before PBD between the two groups (Supplementary Table 2).

Table 2 Univariate analysis of clinical characteristics for PPH

As lower TBIL-post and higher TBIL decay were identified as potential protective factors against PPH as mentioned above, receiver operating characteristic (ROC) curves were plotted to determine the optimal critical value which could serve as reasonable goals for PBD to reduce PPH morbidity (Fig. 1A). Both TBIL-post [area under the curve (AUC) 0.759, 95%CI (confidence interval) 0.672–0.846, P<0.001] and TBIL decay (AUC 0.766, 95%CI 0.676–0.855, P<0.001) showed credibly predictive efficacy. The cut-off value (identified using max Youden index) of TBIL-post was 93.0µmol/L (sensitivity 83.78%, specificity 72.61%) while the cut-off value of TBIL decay was 68.5% (sensitivity 86.49%, specificity 69.43%). TBIL-post ≤ 93.0µmol/L or TBIL decay ≥ 68.5% could significantly distinguish patients who were free of PPH from those who were not.

In multivariate logistic analysis (Fig. 1B), CR-POPF [odds ratio (OR) 4.204, 95%CI 1.531–11.550; P = 0.005], and elevated PT-post (OR 1.650, 95%CI 1.020–2.671; P = 0.041) were identified as significant independent risk factors for PPH. TBIL-post ≤ 93.0µmol/L (OR 0.234, 95%CI 0.064–0.858; P = 0.028) and TBIL decay ≥ 68.5% (OR 0.191, 95%CI 0.048–0.758; P = 0.019) were significant independent protective factors against PPH.

Fig. 1

A. Receiver Operating Characteristic curves illustrated the “diagnostic” performance of TBIL-post and TBIL decay in predicting the onset of PPH at various cutoff values. The AUC for TBIL-post was 0.759 (95%CI 0.672–0.846, P<0.001), and the AUC for TBIL decay was 0.766 (95%CI 0.676–0.855, P<0.001). Based on the max Youden index criterion, the optimal cutoff values were determined to be 93.0µmol/L for TBIL-post and 68.5% for TBIL decay, respectively. B. Multivariate logistic regression analysis identified significant independent risk and protective factors for PPH. CR-POPF and elevated PT-post were identified as significant independent risk factors for PPH. Conversely, TBIL-post ≤ 93.0µmol/L and TBIL decay ≥ 68.5% were significant independent protective factors against PPH

Discovering the optimal duration of preoperative biliary drainage

Excessively prolonging the biliary drainage duration prior to PD has been assumed as a potential cause of surgery delay. In this study, most patients terminated PBD within 6 weeks (177cases, 91.2%), and nearly half of the patients completed PBD within 2 weeks (93cases, 47.9%) (Supplementary Fig. 1).

The dynamic change of TBIL median values and normalized TBIL median values in process of biliary drainage (within 8 weeks) was plotted as curves. A sharp and rapid decline in TBIL median values and normalized TBIL median values was observed in the first 2 weeks, followed by a flattening or even a slight increase as the PBD duration was prolonged (Fig. 2A and B). By the 3rd week, both median TBIL (82.9µmol/L, 64.0µmol/L-133.7µmol/L) and TBIL decay (69.4%, 56.4-79.8%) already met the target values of TBIL ≤ 93.0µmol/L or TBIL decay ≥ 68.5%.

Fig. 2

presented a set of curves illustrating the dynamic change of TBIL relevant data in process of PBD within 8 weeks. A and B displayed the median values of TBIL and normalized TBIL levels following PBD. The median TBIL ranging from 0 to 8 weeks were 277.4 (229.8-349.4) µmol/L, 155.3 (118.9-201.4) µmol/L, 101.1 (79.4-147.7) µmol/L, 82.9 (64.0-133.7) µmol/L, 77.6 (48.2-109.4) µmol/L, 68.5 (42.5–98.7) µmol/L, 61.9 (43.5–89.9) µmol/L, 49.6 (28.4–76.1) µmol/L and 49.9 (43.9-115.7) µmol/L. The median TBIL decay during PBD ranging from 0 to 8 weeks were 0.0%, 43.2 (33.0-56.3) %, 62.9 (49.5–75.5) %, 69.4 (56.4–79.8) %, 73.7 (64.7–82.5) %, 77.0 (60.9–84.9) %, 79.2 (69.4–82.3) %, 82.3 (72.5–88.1) % and 84.1 (59.2–88.4) % separately

The Kaplan-Meier curves were plotted to evaluate the proportion of cases meeting the goals over the duration of PBD (Fig. 3). Among the total 194 cases, the median time to achieve both PBD goals was 3 weeks. The cumulative incidence of achieving TBIL ≤ 93.0µmol/L increased rapidly from 14.4% in the 1st week to 45.4% in the 2nd week and 65.5% in the 3rd week. Similarly, the cumulative incidence of reaching TBIL decay ≥ 68.5% rose from 13.4% in the 1st week to 39.7% in the 2nd week and 59.8% in the 3rd week. However, the growth gradually slowed down thereafter and nearly plateaued after 6 weeks. Notably, over 20% of the patients failed to achieve the goals before terminating PBD and proceeding to surgery, suggesting that simply prolonging the PBD duration to pursue a more thorough relief of jaundice may not be justified.

Fig. 3

Depicted Kaplan Meier curves with adjusted data to present the cumulative incidence of achieving PBD goals as PBD duration prolonged. The median PBD durations of achieving goals were both 3 weeks. The percentage of achieving the goal of TBIL ≤ 93.0µmol/L from 1 to 8 weeks were 14.4%, 45.4%, 65.5%, 70.1%, 75.8%, 77.3%, 78.4% and 78.4% separately. The percentage of achieving the goal of TBIL decay ≥ 68.5% from 1 to 8 weeks were 13.4%, 39.7%, 59.8%, 69.6%, 72.2%, 74.2%, 75.3% and 75.3% separately

Since 3 weeks of PBD may be sufficient for most patients and extending the PBD duration beyond 6 weeks did not appear to offer additional benefits for achieving the goals, the effect of prolonging PBD duration on perioperative safety was compared (Table 3). The incidence of abnormal PBD drainage (including PTCD tube blockage, detachment or biliary stent blockage or displacement) (29.4% vs. 8.5%, P = 0.019) was significantly higher in the group with PBD duration>6 weeks, as was the incidence of incision infections (23.4% vs. 5.6%, P = 0.024). However, no significant differences were observed in intraoperative blood loss, the morbidity of PBD relevant infections, SSI or intra-abdominal infections (P>0.05). It should be noted that it remains challenging to determine whether the higher incidence of abnormal PBD drainage in the duration>6 weeks group was the cause or consequence of the prolonged PBD duration.

Table 3 Comparison of perioperative characteristics in patients with PBD duration more or less than 6 weeks

Comparison of efficacy and safety of EBS versus PTCD in achieving optimal TBIL levels

In overall, the proportion of achieving TBIL-post ≤ 93.0µmol/L (70.8% vs. 51.1%, P = 0.005) and TBIL decay ≥ 68.5% (67.0% vs. 50.0%, P = 0.017) was significantly higher in the EBS group than those in the PTCD group (Table 4). Although no significant differences were found in the dynamic changes of median TBIL level (P>0.05) or median normalized TBIL decay (P>0.05) between EBS and PTCD (Fig. 4A and B), a slight advantage was observed that EBS achieved the goals earlier than PTCD in the 3rd week. Furthermore, Kaplan-Meier analysis indicated that EBS was more conducive in achieving TBIL-post ≤ 93.0µmol/L compared to PTCD (P = 0.002), with the median time of 2weeks in EBS group, compared to 3.5 weeks in PTCD group (Fig. 5A). The median time for achieving TBIL decay ≥ 68.5% was both 3 weeks in EBS and PTCD groups (Fig. 5B). Although PTCD demonstrated comparable efficacy in achieving TBIL decay ≥ 68.5%, EBS showed a short-term advantage within the first 3 weeks.

In terms of PBD relevant safety, the morbidity of hemorrhage, biliary infection and abnormal biliary drainage showed no significant differences between EBS and PTCD (P>0.05) (Table 4). Although patients undergoing EBS faced with the risk of ERCP relevant pancreatitis (14.2% vs. 0.0%, P<0.001), the overall morbidity did not increase compared to PTCD (P>0.05).

Fig. 4

A. Line plot depicted changes in TBIL levels over time. In EBS group, the median levels of TBIL during PBD ranging from 0 to 8 weeks were 271.7 (222.5-348.8) µmol/L, 152.7 (112.2-194.6) µmol/L, 99.0 (69.6-141.1) µmol/L, 74.2 (57.1-125.1) µmol/L, 67.9 (42.6-110.1) µmol/L, 78.4 (49.7-108.7) µmol/L, 61.9 (40.9–89.6) µmol/L, 46.6 (29.4–81.9) µmol/L and 77.8 (45.1-113.5) µmol/L separately. In PTCD group, the median levels of TBIL during PBD ranging from 0 to 8 weeks were 296.4 (237.1–350.0) µmol/L, 158.9 (123.9-209.7) µmol/L, 115.8 (88.9-161.8) µmol/L, 106.9 (73.6-135.7) µmol/L, 86.2 (60.1-109.4) µmol/L, 59.9 (36.0-88.5) µmol/L, 62.6 (48.2–87.3) µmol/L, 49.6 (27.7–56.5) µmol/L and 49.9 (35.0-89.7) µmol/L separately. The two groups showed no significant differences according to Wilcoxon matched-pairs signed rank test (P value = 0.360). B. Line plot depicted changes in normalized TBIL (100%-TBIL decay) over time. In EBS group, the median TBIL decay during PBD ranging from 1 to 8 weeks were 43.0 (32.5–60.1) %, 64.9 (49.3–77.2) %, 69.5 (54.9–81.0) %, 72.4 (57.6–82.9) %, 77.0 (59.5–87.1) %, 79.2 (60.1–83.2) %, 81.8 (69.9–88.1) % and 71.7 (50.9–85.6) % separately. In PTCD group, the median TBIL decay during PBD ranging from 1 to 8 weeks were 43.8 (34.3–51.8) %, 62.6 (50.2–69.7) %, 67.3 (58.6–77.6) %, 75.0 (66.2–82.1) %, 77.1 (66.3–81.6) %, 79.7 (74.1–82.1) %, 86.9 (81.8–87.8) % and 88.4 (76.7–90.5) % separately. The two groups showed no significant differences according to Wilcoxon matched-pairs signed rank test (P value = 0.250)

Fig. 5

Depicted Kaplan Meier curves with adjusted data to present the cumulative incidence of achieving PBD goals with separate methods. As shown in A, for achieving TBIL ≤ 93.0µmol/L, the median time was 2 weeks in EBS and 3.5 weeks in PTCD. In EBS, the percentage of achieving TBIL ≤ 93.0µmol/L from 1 to 8 weeks were 18.9%, 50.9%, 76.4%, 78.3%, 84.0%, 84.9%, 85.8% and 85.8% separately, while 9.1%, 38.6%, 50.0%, 58.0%, 64.8%, 67.0%, 67.0% and 67.0% in PTCD. As shown in B, for achieving TBIL decay ≥ 68.5%, the median time was 3 weeks in EBS and PTCD. In EBS, the percentage of achieving TBIL decay ≥ 68.5% from 1 to 8 weeks were 17.9%, 47.2%, 63.2%, 69.8%, 69.8%, 69.8%, 72.6% and 74.5% separately, while 8.0%, 31.8%, 56.8%, 70.5%, 77.3%, 78.4%, 78.4% and 78.4% in PTCD

Table 4 Comparison of clinical data in patients undergoing EBS and PTCD

Discussion

Watchful waiting for the normal status of TBIL level or liver function during PBD is nether reasonable nor beneficial. This study introduces the novel concept of goal-oriented biliary drainage that PBD targeting either TBIL less than 93.0µmol/L or TBIL decay greater than 68.5% prior to PD can notably reduce the morbidity of PPH. This may help answer the question of “when to stop”, complementing previous studies that addressed the question of “when to start,” making PBD more effective and logical [7].

Distal obstructive jaundice can lead to increased intrahepatic and extrahepatic bile duct pressure, impairing bile synthesis and secretion, which results in cholestatic hepatocyte damage, coagulation dysfunction, endotoxemia, intestinal barrier dysfunction and malnutrition [8, 9]. Dysfunction of liver and coagulation caused by severe hyperbilirubinemia poses threatens to the perioperative safety, including large-volume intraoperative blood loss and PPH [10]. In clinical aspects, surgical accident (including pseudoaneurysm of gastroduodenal artery stump, poor suture of gastrointestinal or pancreatic-intestinal anastomosis and insecure hemostasis of surgical site) along with POPF, intra-abdominal infection, hyperbilirubinemia, advanced age and poor nutritional status are commonly considered high-risk factors for PPH [11]. These factors align with our findings, which indicate that elevated TBIL, prolonged PT and the presence of CR-POPF are independent risk factors for PPH.

PBD is recommended for patients with cholangitis/fever or severe symptomatic jaundice or as candidates for neoadjuvant therapy according to NCCN guidelines. The NCCN guidelines base the indications for PBD on clinical manifestations rather than the level of TBIL. This may be partly because the primary goal of PBD is to alleviate symptoms and potentially reduce the morbidity associated with surgery [12]. Moreover, the relationship between asymptomatic hyperbilirubinemia and surgical outcomes remains unclear and even controversial [13, 14]. A high-volume retrospective study enrolled 1,500 cases reveals that a preoperative TBIL greater than 128µmol/L can accurately predict postoperative complications [15]. Other high-volume retrospective studies, involving 803 and 1,200 cases, proposes the cut-off value of TBIL 222.3µmol/L and 300µmol/L respectively [16, 17]. A most recent retrospective analysis of 290 consecutive cases indicates that PBD in patients with TBIL above 162µmol/L can effectively reduce morbidity of postoperative complications [18]. Additionally, a retrospective and propensity score-matched analysis of 200 cases recommends a routine PBD in patients with TBIL exceeding 250µmol/L so as to reduce overall postoperative complications [7].

It is important to note that the end-point for PBD may differ from the previously reported “start-point” as presented above. Patients candidate for PBD may suffer from more severe symptoms and hepatocyte damage, and the primary tumor may continue to progress during the procedure, compounded by the toxic effects of persistent hyperbilirubinemia. Therefore, the “end-point” value is necessary to be set at a relatively lower level to provide comprehensive liver function relief. For proximal obstructive jaundice caused by hilar cholangiocarcinoma, a cut-off value of TBIL ≤ 75µmol/L has been reported to reduce intraoperative blood loss and conducive for short-term survival after hepatectomy, with earlier studies recommending the proper TBIL level as low as 34–51µmol/L [19]. However, for distal obstructive jaundice, our research identified a TBIL level of 93.0µmol/L prior to PD as the optimal cut-off value based on the maximum Youden index in the ROC curves. Subsequent multivariate regression analysis confirmed that TBIL ≤ 93.0µmol/L is an independent protective factor which may whittle down the risk of PPH by more than 70%. In order to provide a more flexible recommendation of the PBD end-point, we propose a “relativity” indicator: TBIL decay ≥ 68.5%. This indicator offers predictive and protective value comparable to the absolute TBIL value, as it accounts for different baseline TBIL levels. The equivalent percentage of TBIL decay in higher hyperbilirubinemia implies a greater alleviation of the jaundice burden.

Prolonging PBD unnecessarily may delay the opportunity for surgery. We observed a sharp decline in median TBIL level within the first 2 weeks, followed by a plateau or even a reverse upward tendency when PBD extends beyond 6 weeks. This may be attributed to the decreasing siphon effect as biliary pressure diminishes. Consequently, we are highlighting to set an alarm clock at the 3rd week as a check-point, unless the patient is receiving neoadjuvant therapy [20]. It has been reported that 6 weeks can be generally sufficient to normalize the liver functions with complete biliary drainage [21]. According to the European Society of Gastrointestinal Endoscopy’s clinical guidelines, a TBIL decay less than 20% from baseline on the seventh day after PBD is indicative of inadequate drainage [22]. Additionally, on occasion that the jaundice status is still far from the goal after 6 weeks, the possibility of an improper initial placement of the stent or PTCD tube should be reconsidered. It is also important to check for abnormal drainage events including tube blockage, detachment or stent occlusion or migration. The preference for plastic stents, commonly used for temporary biliary drainage, is concerning as their patency rates drop to 48.1% and 26.8% after 8 and 12 weeks respectively [23]. Previous retrospective studies have revealed that an over 4-week PBD duration may lead to more overall complications [24, 25]. As a drain or stent in the biliary tract may significantly promote the microbial colonization of bile, prolonged PBD is associated with a significant higher risk of infectious complications [26]. A recent retrospective study indicated that the risk of incision infection increases after 4 to 6weeks post-stenting [27], aligning with our findings that PBD beyond six weeks significantly elevates the incidence of incision infections. This suggests that careful planning is needed to manage PBD duration and optimize the use of perioperative antimicrobial agents to prevent infections [28]. In cases where prolonged PBD is necessary due to neoadjuvant therapies, extended antibiotic prophylaxis should be considered to protect against SSI, especially for patients with EBS, without a significant risk of increasing short-term bacterial resistance [29, 30].

Selecting the appropriate PBD method may be more conducive to achieving the goals. The placement of self-expanding metal stents (SEMS) is recommended by NCCN guidelines because they are friendly to place without dilation and have less impact on subsequent resection and maintain a longer patency rate. Our observations indicate that EBS is generally more effective in achieving the goals compared to PTCD. Furthermore, we have also found that EBS and PTCD have similar TBIL decay though PTCD appears to have a shorter PBD duration than EBS, albeit without statistical significance. However, it should be noted that the comparable efficacy of EBD and PTCD is mainly observed in short-term biliary drainage bridging to surgery, in which both methods may provide equivalently good patency. In cases requiring longer PBD durations, particularly those involving neoadjuvant therapy, EBS may offer superior performance. Although pancreatitis is a complication unique to EBS, the incidence of overall PBD relevant complications is similar for both methods. While PTCD has traditionally been an effective rescue therapy for patients who fail EBS or experience abnormal drainage events [31], recently studies have validated endoscopic ultrasound-guided stenting as a viable alternative [32].

It is important to emphasize that the goal of PBD is not to simply address hyperbilirubinemia itself. The purpose of PBD is to alleviate obstructive jaundice, improve liver and coagulation function, and optimize clinical conditions for subsequent surgery and postoperative recovery. The targets of TBIL-post or TBIL decay are proposed to reflect the degree of jaundice relief and their association with observed clinical outcomes. To some extent, improvement of clinical symptoms, such as vomiting, pruritus and coagulopathy are considered more important than a simple reduction in TBIL levels. In this context, goal-oriented PBD not only helps to reduce the risk of PPH, but also avoids unnecessarily prolonging the bridging period and delaying surgery. With this approach, we continually remind surgical decision-makers that “don’t hesitate to schedule for surgery at any time proper”. We are also hoping that this concept will lead to more precise PBD management and provide insights for answering the question of “when to stop”.

Limitations

Patients in this study may undergo varying diagnostic and treatment modalities over the 7-year study period due to advancements in medical techniques. These variations may have affected the outcomes for patients treated at different points during the study. Additionally, this investigation was retrospective and conducted at a single center, meaning that the inherent biases of such a design cannot be entirely excluded.