A systematic review of the comparison of the incidence of seeding metastasis between endoscopic biliary drainage and percutaneous transhepatic biliary drainage for resectable malignant biliary obstruction

Patients diagnosed with perihilar cholangiocarcinoma (PHC), distal cholangiocarcinoma (DCC), and pancreatic cancer (PC) typically present with malignant biliary obstruction (MBO), which is one of the crucial reasons for the failure of surgery [1]. Preoperative biliary drainage (PBD) is deemed to improve jaundice before surgery and decrease postoperative morbidity and mortality, although it remains controversial [2‐4]. Furthermore, either percutaneous transhepatic biliary drainage (PTBD) or endoscopic biliary drainage (EBD) is the best strategy for resectable MBO is also a question [5‐9].

Seeding metastasis (SM) is rarely refereed worldwide, but it has been reported frequently in Japan [10‐17]. The incidence of SM in Japan was reported to range from 4.0 to 40.4% [10‐17], which is no longer an “unusual” contingency. EBD was reported repeatedly to superior to PTBD in the prophylaxis of SM [10‐17], but it was contradicted by a multicenter, retrospective study derived from US-Euro [18]. Hence, a systematic review is warranted to evaluate the incidence of SM between EBD and PTBD in the procedure of PBD for patients with MBO.

PHC, DCC, and PC are typically present with biliary obstruction, which often increases the risk of perioperative mortality and morbidity [21, 22]. Palliative biliary drainage has been repeatedly confirmed to be efficient in the improvement of prognosis for patients with unresectable MBO [23, 24], but it still remains controversial whether patients with resectable MBO would be benefited from PBD [2‐4]. Furthermore, whether either PTBD or EBD is better is another puzzle [5‐9], although both have been conducted prevalently worldwide. Short-term outcomes, such as pancreatitis, bile leakage, and clinical and technique success rates, are the common indicators to compare the efficacy and safety between PTBD and EBD. From this aspect, EBD has been confirmed superior to PTBD by several meta-analyses [25‐27]. Long-term outcome of PBD was rarely taken into consideration of the strategy for patients with MBO, but the superiority of EBD in the overall survival was reconfirmed in our previous meta-analysis [28].

SM is rarely reported globally, but it was reported as high as 4.0~40.4% in Japan [10‐17]. Reasons are as follows: (a) preoperative PTBD longer than 60 days was associated with an increased risk of the SM [29], (b) repeated attempts at PTBD [29, 30], and (c) multiple plastic stents were used rather than single one [29, 30]. In this study, eight included studies came from Japan [10‐17], and the total incidence of SM (18.4%) was higher than that from Korea [19] (2.0%) and comparable with Netherland-USA [18] (16.7%). In the view of statistical data, SM was no longer an incident. Hence, the issue of SM deserved much more attention in clinical.

Theoretically, EBD is unlikely to cause SM. In this study, the incidence of SM is much lower in the EBD group than that in the PTBD group (10.5% vs. 22.0%, P < 0.001). Potential mechanism of increased PTBD-related SM lied in that as follows: (1) tumor cells derived from the PTBD fluid drainage were reported to be more than those from the EBD [19], which indicated that PTBD was much more likely to cause tumor diffusion. (2) PTBD was performed conventionally in the right liver, where the liver, peritoneum, diaphragm, pleura, and subcutaneous tissue were very close to each other anatomically [17], which offered an appropriate environment for SM. Hence, Takahashi et al. [29] recommended that PTBD should be performed in the left. (3) The general condition was usually much poorer in the patient intended to conduct PTBD, which meant a higher risk for SM.

The location of the MBO might be taken into consideration to decide an appropriate PBD. The recommended level of PTBD is higher than that of EBD for PHC according to the Chinese guideline (IIA vs. IIB) [31], but it remains controversial for DCC and PC in most of the guidelines [5, 6, 8]. In this study, the incidences of SM differed significantly among PHC, DCC, and PC (16.2% vs. 11.6%, and 22.5%), partly owing to varied aggressive characteristics of different cancers. However, subgroup analyses which were stratified by PHC, DCC, and PHC showed that the incidences of SM were lower in the EBD groups than those in the PTBD groups (7.8% vs. 11.7%, P < 0.001; 6.0% vs. 18.2%, P < 0.001; 15.6% vs. 40.3%, P < 0.001; respectively), which indicated that EBD was superior to PTBD in the prevention of SM regardless of the location of MBO.

However, we felt puzzled for the mechanism of SM caused by EBD. Technically, EBD was unlikely to cause peritoneal metastasis unless intestinal perforation occurred in the procedure of EBD, especially in the experienced centers. In this study, the overall incidence of tube-related SM in the EBD group is 2.0%. Theoretically, EBD was unlikely to cause peritoneal SM, because the whole procedure of EBD was conducted inside the biliary duct. In this study, we divided SM into tube-related SM and peritoneal SM, and subgroup analysis showed that both the rate of tube-related SM and peritoneal metastasis decreased significantly in the EBD group (2.0% vs. 6.7%, P < 0.001; 10.0% vs. 20.2%, P < 0.001; respectively). Hence, we concluded that EBD was superior to PTBD in the prophylaxis of SM, although the definition of SM remained controversial.

There were several limitations in this study. First, there were no RCTs included in this meta-analysis, which made the conclusion sound weaken because cohort data had selection bias. Second, studies included in this meta-analysis were nearly from Japan, which indicated obvious regional bias. Third, PTBD was available when EBD failed, but those patients belonged to which group remained inconsistent [15, 19]. Bias could be also due to the following: (a) the requirement for an alternative drainage procedure due to therapeutic or technical failures was likely higher in the EBD group compared to patients undergoing preoperative PTBD and (b) Bismuth type III and IV tumors as compared to type I and II tumors were potentially better decompressed percutaneously, especially in the presence of complex strictures. Fourth, the definition of SM varied from each other due to the lack of a golden standard, for example, intrahepatic metastasis belonged to PBD-related seeding metastasis in PC [14], but as for PHC and DCC, it tended to be rich in contradiction. Fifth, technical parameters of either PTBD or EBD, such as the procedure of PBD and the stent material, were different from each center, which indicated an inevitable heterogeneity and weaken the reliability. Sixth, the severity of obstructive jaundice, i.e., the level of preoperative serum bilirubin, was far from consistency, and caution should be taken when interpreting these results. The last but not the least, it was hard to avoid publication bias, because the journals tend to publish positive results.

In summary, we concluded that EBD was superior to PTBD for resectable MBO in the prophylaxis of SM, but there were currently not enough evidences. In the future, working out the definition of SM related to PBD is the urgent affair.