Background

Endoscopic retrograde cholangiopancreatography (ERCP) with transpapillary stenting has become the gold standard procedure for biliary decompression in patients with obstructive jaundice, due to its high efficacy and low morbidity [1]. However, even in expert hands, ERCP fails in 3–12 % of cases [2]. One of the most common causes of ERCP failure is duodenal stenosis, responsible for around 16 % of failed ERCP [3]. In these cases, alternative approaches include surgical intervention, percutaneous transhepatic drainage (PTD) and, more recently, endoscopic ultrasound (EUS)-guided biliary drainage, all of which carry higher morbidity than ERCP [4–6].

Duodenal covered self-expandable metal stents (cSEMSs) were developed to prevent tumor ingrowth in malignant gastric outlet obstruction [7]. As cSEMSs are removable, they offer a promising alternative to surgery or failed endoscopic dilation for benign mechanical obstructions such as peptic ulcer, chronic pancreatitis or radiation-induced stenosis [8, 9].

In cases of failed ERCP due to duodenal stenosis, temporary cSEMS placement could usefully provide secondary access to the ampulla of Vater and potentially achieve retrograde biliary drainage. However, the literature has not yet addressed the use of cSEMS in this indication.

Here we report a study evaluating the safety and efficacy of temporary cSEMS placement to access the papilla and achieve secondary retrograde biliary drainage in patients with duodenal stenosis in whom ERCP had initially failed.

Methods

Patients and study design

This study included all patients presenting to our referral center with obstructive jaundice without severe sepsis and who underwent a temporary cSEMS placement after a failed ERCP due to duodenal stenosis during a 9-year period. Procedures were prospectively collected in this observational, single-center study. The cSEMS was placed at the same time as the failed ERCP, and left in place for 7 days, during which the following clinical and biological criteria were assessed: fever, nausea/vomiting, hemorrhage, pain, lipasemia, bilirubinemia at baseline and at day 7, complete blood count and biochemical tests of renal function. Conventional radiography was performed before the second ERCP to assess for stent migration. Antimicrobial agents were only given in cases presenting clinical signs of acute cholangitis according to the Tokyo guidelines [10]. After 7 days, cSEMSs were removed and access to the papilla was recorded. When the papilla was accessible by the duodenoscope, a retrograde biliary drainage was attempted. In cases of second failed ERCP, either EUS-guided biliary drainage or percutaneous transhepatic drainage was performed. Primary endpoint was papilla accessibility rate at 7 days post-cSEMS placement. Secondary endpoints were ERCP success rate at 7 days post-placement, bilirubinemia at baseline versus day 7, technical success, stent migration rate and other adverse events such as cholangitis, pancreatitis, gastrointestinal bleeding, perforation or renal failure according to ASGE lexicon.

The study was performed in accordance with the Declaration of Helsinki, good clinical practice and all applicable regulatory requirements. This clinical trial was approved by the local Institutional Review Board (Clermont-Ferrand Hospitals IRB #00008526/Ref: 2015/CE114). Written informed consent was obtained from all patients prior to endoscopic procedures.

Endoscopic procedure

All procedures were performed by experienced operators and with patients in supine position, intubated and sedated with propofol, in an interventional endoscopy room, under endoscopic and fluoroscopic guidance. Endoscopes used were a sideviewing duodenoscope with a 4.2-mm working channel (TJF 160, Olympus Tokyo, Japan), a colonoscope with a 3.8-mm working channel (AOI 180, Olympus) or a 2-channel gastroscope with a 3.7-mm working channel (GIF-2T200, Olympus). Through-the-scope (TTS) silicon-covered nitinol Hanarostents (NDC Duodenum/Pylorus or DPC Duodenum/Pylorus, M.I. Tech, Korea) were used in all patients. In a first step, 50 cm3 of Telebrix 30 was injected via the working channel of the endoscope to assess stenosis length and morphology. A 0.035-inch guidewire (Dreamwire, Boston Scientific) was passed through the stenosis using an ERCP catheter (Tandem XL 5.5F, Boston Scientific) and advanced to the angle of Treitz. If the gastroduodenal stenosis was difficult to catheterize, the patient was placed in left lateral decubitus to facilitate catheterization. When initial opacification was unable to assess stenosis length and locate the distal end of the stenosis because it was too tight, opacification had to be performed via the triple lumen catheter advanced over the guidewire. If required, the wire was swapped to an extra-stiff 430-cm guidewire (Wallstent, Boston Scientific) to enable insertion of the delivery device if the duodenal anatomy presented angulated distortion. The stent delivery system was then advanced over the guidewire through the stricture, under fluoroscopic guidance. The stent was at least 4 cm longer than the stenosis (2 cm longer at each end) in order to cover the entire stenosis and anticipate post-procedure stent shortening due to continued radial stent expansion. The stent was deployed from distal to proximal tip, through the pylorus, with a length of nearly 2 cm in the gastric antrum, under fluoroscopic and endoscopic control. Balloon dilation before or after stent placement was not performed. Adequacy of stent placement and stent performance was assessed post-deployment by endoscopy and fluoroscopy using a contrast injection through the scope. The stent was removed 7 days after placement using either a large snare surrounding its proximal tip in the stomach or rat-tooth forceps by grasping the retrieval lasso when DPC stents were used. At this point, we simultaneously attempted to gain access to the papilla with the duodenoscope. Crossing duodenal stenosis was systematically attempted without balloon dilation. When access to papilla was possible, retrograde biliary drainage was achieved whenever biliary cannulation through guidewire insertion was effective.

Statistical analysis

Data were collected and managed using REDCap electronic data capture tools hosted at Clermont-Ferrand University Hospital. Baseline characteristics and results are expressed as mean ± standard deviation with 95 % confidence intervals (CIs). Tests for correlations between outcome of retrograde biliary drainage and population characteristics used a Chi-squared test (or Fisher’s exact test when necessary) for categorical variables and the Student’s t test (or Kruskal–Wallis test) for quantitative variables. A p value of <0.05 (two sided) was considered significant. Calculations were performed using Stata 12 (StataCorp, College Station, TX).

Results

From June 2006 to March 2014, 281 patients underwent 367 duodenal SEMS placement procedures, of which 154 were cSEMS. Of these 154 cSEMSs, 26 patients (68 % male, mean age 69 ± 13 years) presented with obstructive jaundice without severe sepsis and underwent a temporary 7-day cSEMS during an initial failed ERCP due to an impassable duodenal stenosis. Fourteen patients (54 %) presented with obstructive symptoms, i.e., nausea, vomiting, bloating with no or liquid-only intake [score ≤ 1 on the gastric outlet obstruction scoring system (GOOSS)]. Three patients (12 %) presented with initial cholangitis and were treated with intravenous antibiotics. Duodenal stenosis and jaundice were due to malignant disease in 21 (81 %) cases and benign diseases in 5 (19 %) cases. Patient characteristics are reported in Table 1. Duodenal cSEMS placement was technically successful in all cases, and the 7-day period between the two procedures was marked by zero complications such as cholangitis, pancreatitis, gastrointestinal bleeding, perforation or renal failure. Proximal stent migration in the stomach occurred in 6 cases (23 %): 5 patients remained asymptomatic and one presented with recurrent obstructive symptoms. Mean bilirubin level (±SD) was 102 ± 90 µmol/L at baseline rising to 164 ± 121 µmol/L at day 7. At day 7, all cSEMSs were successfully retrieved without any cases of indwelling stent or stent fracture. Access to the papilla was possible in 25 cases (96 %, 95 % CI 80–99 %). The only patient in which access to the papilla was impossible was treated by EUS-guided biliary drainage. Once access to the papilla was achieved, secondary retrograde biliary drainage was successfully achieved in 19 cases (76 %, 95 % CI 55–91 %), i.e., 73 % (95 % CI 73–86 %) in an intention-to-treat analysis (Table 2). Five of 6 patients in whom retrograde biliary stenting failed presented tumoral invasion of the papilla. Of these 6 patients, 3 were treated with EUS-guided biliary drainage, one was treated by percutaneous transhepatic drainage, and one patient died as anterograde drainage was contraindicated due to advanced peritoneal carcinomatosis with ascites. Note that biliary drainage was not performed in one patient who presented a spontaneous decrease in bilirubin after acute necrotizing pancreatitis. Table 3 gives patient characteristics according to outcome of retrograde biliary drainage when access to the papilla was achieved. Presence of initial digestive occlusive syndrome was significantly associated with failure of retrograde biliary drainage [0 % (0/11) failed biliary cannulation when patients presented with no GOO symptoms vs 42 % (6/14) in cases with GOO symptoms; p = 0.02]. Among 14 patients who presented with initial obstructive symptoms, 8 (57 %) required a placement of a second duodenal stent due to persistent evidence of GOO, 4 with uncovered SEMS and 4 with covered SEMS.

Table 1 Baseline characteristics of the study population
Full size table
Table 2 Outcomes of patients presenting obstructive jaundice with failed ERCP due to duodenal invasion and treated by temporary cSEMS placement
Full size table
Table 3 Univariate analysis of risk factor for secondary failed ERCP when access to the papilla was achieved (N = 25)
Full size table

Discussion

Simultaneous duodenal and biliary stricture is a very common situation in patients with pancreatic and biliary malignancies [1, 11] as well as in chronic or acute pancreatitis in case of paraduodenal pancreatitis. This situation is one of the most common causes of ERCP failure [12]. Endoscopic dilation to pass the duodenal stenosis has low efficacy [13], particularly in tumor stenosis. In these cases, conventional alternative approaches include percutaneous transhepatic drainage (PTD) and surgical biliodigestive derivations but also carry significant associated morbidity [4, 5]. EUS-guided biliary drainage has recently emerged as an effective therapy for bile duct decompression when ERCP has failed [14] or in patients with an indwelling duodenal stent [15]. It offers the advantage of immediate drainage during the same endoscopic session as the failed ERCP, obviating a second deferred procedure with a radiologist and the risks associated with a second anesthesia, but it also has significantly higher morbidity than ERCP, even if this morbidity rate appears to be decreasing with the learning curve [3]. Moreover, EUS-guided biliary drainage is currently restricted to experienced operators in tertiary care centers. A further limitation to anterograde drainage techniques is that they are contraindicated in cases of ascites due to a significant increase of risk of bile leakage [6]. In these situations, patients often reach a therapeutic impasse.

Uncovered duodenal SEMS are widely used in patients with advanced malignant gastroduodenal obstruction [16]. Covered duodenal SEMS (cSEMS) is also an option in malignant disease settings to reduce tumor ingrowth or hyperplasia, as well as in patients scheduled for surgery or in emergency procedures involving unknown resectable status.

This is the first study to report on experience with temporary cSEMS placement after initial ERCP failure due to duodenal stenosis to achieve access to the papilla and secondary successful ERCP. Indeed, to our knowledge, the use of duodenal cSEMS in this indication has not yet been described in the literature. Technical success was achieved in all cases with no stent incarceration or stent fracture at stent removal, in contrast with several reported series on the use of covered duodenal stents [17, 18]. The absence of stent fracture could be due to the short stent indwell period here (7 days) versus 2 to 7 weeks in other reported series [19, 20]. Stent migration rate was 23 % (6/26), which is consistent with the published series [21]. Moreover, there was no enteral injury as all stent migrations were proximal in the stomach, making the gastric cSEMS easy to remove, with only one case involving persistent obstructive symptoms. After stent removal, access to the papilla was achieved in 96 % (25/26) of cases. This suggests that a 7-day cSEMS placement is a long enough delay for calibration of the duodenal stenosis and to allow duodenoscope passage to the papilla. This delay, although arbitrarily chosen, seemed logical to allow effective duodenal dilation while avoiding the risk of organ failure, malnutrition and sepsis due to prolonged jaundice. We have shown in a previous study that a cSEMS bridging the papilla does not interrupt bile flow from the ampulla of Vater [9]. Furthermore, during the 7-day period between the two ERCPs was produced a median 62 µmol/L increase in bilirubinemia with no cases of complications such as cholangitis, pancreatitis, kidney or hepatic failure. The low median increase of bilirubinemia in this patient population presenting malignant duodenal stenosis in 85 % of cases confirms the absence of stent-induced bile obstruction. This increase in bilirubinemia corresponds to the increase expected after 7 days in cases of tumoral biliary obstruction. This 7-day delay thus appears safe in terms of sepsis or organ failure, and in practice corresponds roughly to the delay required to address a patient to a tertiary care center to get a second anterograde procedure by either EUS-guided biliary drainage or PTD at medical centers that do not perform therapeutic EUS. cSEMS placement thus emerges as a simple and reproducible technique available to all endoscopy units, which is a crucial advantage when selecting a first-line technique in cases of failed ERCP due to duodenal stenosis.

Successful retrograde biliary drainage was achieved in 76 % of cases here without any complications. We know that in an expert’s hands, the failure rate of biliary cannulation is expected to be less than 10 % percent excluding duodenal invasion. Here, 21 out of 25 patients had tumoral obstructive jaundice with periampullary tumoral infiltration and 3 had chronic pancreatitis. Biliary cannulation is more difficult in cases of periampullary tumoral infiltration and can also be challenging in cases of chronic pancreatitis due to distortion of the papilla area. This may explain the higher rate of biliary cannulation failure (24 %) when papilla was accessible during the second ERCP, whereas the rate of failed ERCP for biliary decompression was 9.9 % and the rate of failed cannulation was 4.1 % during the same period in our center [3]. Furthermore, we found that presence of symptoms of gastric outlet obstruction (GOO) during the first ERCP was predictive of failed biliary cannulation at secondary access to the papilla (Table 3), and more than half of the patients here (14/26) presented with obstructive syndrome. GOO symptoms could be related to a more advanced malignant gastroduodenal obstruction, and thus unresectable distal biliary obstruction [15].

Finally, this temporary cSEMS strategy avoids a riskier anterograde biliary drainage in almost three-fourth of cases in patients with failed ERCP due to an impassable duodenal stenosis. Nevertheless, the technique should be reserved to patients without severe sepsis or initial renal failure, as these situations warrant urgent biliary drainage and thus an alternative approach.

Limitations to this study include the fact that it was performed at a single center, with a relatively small number of patients, and a retrospective analysis of a prospectively collected database. A larger multicenter prospective randomized study is now needed to compare the results of delayed ERCP after temporary duodenal cSEMS against anterograde biliary drainage and determine the role of this technique as a first-line alternative therapy when ERCP fails due to duodenal invasion.

In conclusion, in 7-day temporary placement of a cSEMS in cases of obstructive jaundice with failed ERCP due to duodenal stenosis was found to be a safe and effective strategy to access the papilla, leading to successful second ERCP in nearly three out of four cases. This approach, which has never previously been reported, can avoid EUS-guided biliary drainage or PTD, both of which carry higher morbidity.