- Email: [email protected]
No Benefit of Endoscopic Sphincterotomy Before Biliary Placement of Self-Expandable Metal Stents for Unresectable Pancreatic Cancer
Clinical Gastroenterology and Hepatology 2015;13:1151–1158
PANCREAS, BILIARY TRACT, AND LIVER No Benefit of Endoscopic Sphincterotomy Before Biliary Placement of Self-Expandable Metal Stents for Unresectable Pancreatic Cancer Tsuyoshi Hayashi,* Hiroshi Kawakami,‡ Manabu Osanai,§ Hirotoshi Ishiwatari,* Hirohito Naruse,k Hiroyuki Hisai,¶ Nobuyuki Yanagawa,# Hiroyuki Kaneto,** Kazuya Koizumi,‡‡ Tamaki Sakurai,§§ and Tomoko Sonodakk *Department of Medical Oncology and Hematology, kkDepartment of Public Health, Sapporo Medical University School of Medicine, Sapporo, Japan; ‡Department of Gastroenterology and Hepatology, Hokkaido University Graduate School of Medicine, Sapporo, Japan; §Center for Gastroenterology, Teine-Keijinkai Hospital, Sapporo, Japan; kDepartment of Gastroenterology, Hakodate Municipal Hospital, Hakodate, Japan; ¶Department of Gastroenterology, Japan Red Cross Date General Hospital, Date, Japan; #Department of Gastroenterology, Asahikawa-Kosei General Hospital, Asahikawa, Japan; **Department of Gastroenterology, Muroran City General Hospital, Muroran, Japan; ‡‡Division of Gastroenterology and Hematology/Oncology, Department of Medicine, Asahikawa Medical University, Asahikawa, Japan; §§Department of Gastroenterology, Steel Memorial Muroran Hospital, Muroran, Japan BACKGROUND & AIMS:
Endoscopic sphincterotomy (ES) is performed routinely before self-expandable metallic stents (SEMS) are placed in malignant distal biliary strictures to prevent postprocedural pancreatitis. However, it is not clear whether ES actually prevents pancreatitis or affects other adverse events (AEs). We conducted a noninferiority trial to examine the necessity of ES before SEMS placement.
METHODS:
Two hundred patients with distal biliary strictures caused by unresectable pancreatic cancer were assigned randomly to groups that received ES or did not receive ES (non-ES) before SEMS placement, at 25 hospitals in Hokkaido, Japan, from August 2010 through November 2012. The primary outcome was early AEs (£30 d) specifically related to the presence or absence of ES (pancreatitis, bleeding, or perforation). Secondary outcomes measured included the effect of ES omission on time to SEMS dysfunction and patient survival times.
RESULTS:
The proportions of patients with early AEs were 9.2% in the non-ES group and 10.4% in the ES group (a difference of 1.2%, noninferior). The median times to SEMS dysfunction was longer than 594 days in the non-ES group and 541 days in the ES group (P [ .88). The median overall survival times were 202 in the ES group vs 255 days in the non-ES group; P [ .20).
CONCLUSIONS:
ES before SEMS does not affect the incidence of AEs, SEMS patency, or patient survival times. Our data provide no evidence for a benefit of ES to patients undergoing SEMS placement for a biliary stricture caused by pancreatic cancer. UMIN clinical trials registry number: 000004044.
Keywords: Endoscopic Sphincterotomy; Self-Expandable Metallic Stent; Adverse Event; Pancreas; Distal Biliary Stricture.
See editorial on page 1159. istal biliary stricture is a common comorbidity of unresectable pancreatic cancer. Because resection is not feasible at the time of diagnosis for one third of pancreatic cancers,1 palliation of biliary stricture is important to improve quality of life and allow initiation of chemotherapy. Decompression of biliary strictures usually is accomplished by endoscopic stent placement. Self-expandable metallic stent (SEMS) placement is now a standard procedure for distal biliary stricture caused
D
by unresectable malignant lesions and achieves longer stent patency than plastic stent placement because of the larger diameter of the SEMS.2–5 However, pancreatitis is a common postprocedural complication. Many endoscopists believe that this adverse event (AE) can Abbreviations used in this paper: AE, adverse event; CI, confidence interval; ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic sphincterotomy; ITT, intention-to-treat; SEMS, self-expandable metallic stent. © 2015 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2015.01.008
1152 Hayashi et al
be prevented by endoscopic sphincterotomy (ES). To date, only 1 prospective randomized controlled trial has examined the influence of ES on the rate of AEs associated with SEMS placement for malignant biliary stricture. They concluded that ES is disadvantageous in patients with malignant biliary stricture based on the lower rate of SEMS-related AEs in the non-ES group compared with the ES group (11% vs 49%; P ¼ .006).6 However, the combined rate of bleeding and perforation in the ES group was 24.3%,6 which was exceedingly high relative to other reports of SEMS placement with ES (0%–1%).2–5,7–11 Furthermore, post–SEMS placement pancreatitis was not observed in either the non-ES or ES groups.6 This is in contrast to several previous studies that have reported that pancreatitis is the most frequent early post-SEMS placement AE (0%–8.8%).2–5,7–15 Because these results do not reflect the typical clinical practice experience and impaired pancreatic outflow is considered a major cause of pancreatitis, the majority of endoscopists continue to perform ES routinely before SEMS placement to minimize the influence of the strong expansive force of the SEMS on the pancreatic duct orifice. In Japan, some endoscopists believe ES before SEMS placement is useless, especially in pancreatic cancer patients in whom exocrine function already has ceased because of a main pancreatic duct obstruction. Therefore, we conducted a multicenter, prospective, randomized, controlled trial to investigate the influence of ES on the incidence of AEs associated with SEMS placement for biliary stricture in patients with unresectable pancreatic cancer (the multicenter, prospective, randomized, controlled trial to investigate the influence of endoscopic sphincterotomy on the incidence of adverse events associated with self-expandable metallic stent placement for biliary stricture in patients with unresectable pancreatic cancer trial in Hokkaido).
Patients and Methods Study Design Twenty-five hospitals in Hokkaido, Japan, participated in the study including 3 educational, 5 tertiary referral, and 17 secondary referral hospitals. The treatment (non-ES or ES) was assigned randomly upon enrollment by a 1:1 centralized allocation using the minimization method. The allocation and data collection were performed using a web-based system that was unavailable to the investigators. Randomization was concealed from patients, endoscopists, and physicians. However, patients, endoscopists, and physicians were unmasked to the treatment allocation. The study protocol was approved by the Institutional Review Board of each participating hospital and registered with the University Hospital Medical Information Network Clinical Trials Registry (http://www.umin.ac.jp/ctr/index/htm/; no. University Hospital Medical Information Network
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
000004044). All authors had access to the study data and reviewed and approved the final manuscript.
Patients Complete medical history, physical examination, Eastern Cooperative Oncology Group performance status, blood examination, contrast enhanced-computed tomography and/or magnetic resonance imaging study findings, endoscopic examinations, and cytohistologic diagnosis were used to assess the eligibility of patients with distal biliary stricture caused by unresectable pancreatic cancer. Patients meeting all of the inclusion and none of the exclusion criteria were considered eligible (Table 1). Written informed consent to participate in the study was obtained from all participants at each hospital before randomization.
Self-Expandable Metallic Stent Placement SEMS placement was performed as an inpatient procedure in the participating hospitals, each of which had at least 1 experienced endoscopist (defined as 10 years of postresidency experience, 2000 endoscopic retrograde cholangiopancreatographies (ERCPs), or an annual ERCP caseload >200). All units permitted trainees to perform SEMS placement under the supervision of an experienced endoscopist. All patients Table 1. Eligibility Criteria Inclusion criteria Distal biliary stricture resulting from unresectable pancreatic cancer diagnosed based on contrast-enhanced computed tomography and/or magnetic resonance imaging findings Histologically or cytologically confirmed pancreatic cancer Age 20 y Ability to understand and willingness to sign a written statement of informed consent Exclusion criteria Moderate or severe cholangitis according to the Tokyo guidelines of cholangitis18 Involvement of the ampulla of Vater confirmed by endoscopy Billroth II or Roux-en-Y reconstruction after gastrectomy Intestinal obstruction toward the anal direction of the ampulla of Vater Pancreaticobiliary maljunction Biliary stricture extending to the hilum History of placement of a transpapillary biliary drainage tube with a bore larger than 8F History of previous endoscopic sphincterotomy or precut sphincterotomy Current use of anticoagulant or antiplatelet drugs Prothrombin time–international normalized ratio 1.5 and/or 50,000 platelets/mL ECOG performance status of 4 Severe heart or pulmonary disease Any disorder that compromised the patient’s ability to provide written informed consent and/or comply with the study procedure
ECOG, Eastern Cooperative Oncology Group.
June 2015
underwent SEMS placement under sedation using a combination of diazepam, midazolam, and/or pethidine hydrochloride. Various types of ERCP catheters compatible with the 0.035-inch hydrophilic guidewire were used. The rendezvous method of biliary cannulation was permitted in patients who underwent percutaneous biliary drainage; however, precut sphincterotomy was absolutely prohibited, even for difficult biliary cannulations, in both the non-ES and ES groups. Pancreatic sphincterotomy and prophylactic pancreatic stent placement also were prohibited. In the ES group, after successful guidewire placement and cholangiography, an incision was made using a standard traction-type sphincterotome with a blended current and was continued until visual recognition of bile outflow or before reaching the upper edge of the oral protrusion in cases without outflow. In both groups, a 10-mm diameter WallFlex Biliary RX Partially Covered Stent (Boston Scientific Japan, Tokyo, Japan) was inserted over the guidewire under fluoroscopic guidance. A 40-, 60-, or 80-mm stent was chosen according to the location and length of the biliary stricture as estimated by withdrawing the catheter or guidewire from the proximal end of the stricture to the ampulla of Vater. The SEMS routinely was placed across the papilla with approximately 10 mm of its distal end exposed to the duodenal lumen.
Follow-Up Evaluation After the procedure, all patients fasted until the attending physician confirmed that no AEs had occurred. A blood examination routinely was performed 24 hours after the procedure. Patients who experienced AEs were treated in the hospital as necessary. After discharge, blood parameters, clinical signs, and symptoms of the patients were monitored by a physician at an outpatient clinic at least monthly until patient death or the end of the study period, and all AEs, cholangitis incidence, and patient survival were recorded by the participating hospitals. All clinical information was collected consistently, even for patients who visited or were transferred to affiliated hospitals. Patients who developed recurrent cholangitis were re-admitted to the participating hospital, and endoscopy was performed to investigate the cause and for palliative treatment. The follow-up period ended 6 months after enrollment of the last patient.
Definitions The endoscopic procedure time was defined as the time from guidewire placement to confirmation of successful SEMS placement by visual recognition of bile outflow from the distal end of the SEMS by duodenoscopy or disappearance of the contrast medium from the common bile duct by fluoroscopy. Early and late AEs were defined as any SEMS placement–related AE
Sphincterotomy Before Stent Placement 1153
occurring within 30 days or more than 31 days after SEMS placement, respectively. AEs were defined as follows: (1) post–SEMS placement pancreatitis diagnosed according to the Cotton et al16 criteria (ie, new or worsening upper abdominal pain lasting >24 hours after the procedure in conjunction with hyperamylasemia >3 times the upper limit of the normal range); (2) bleeding diagnosed by the presence of melena or hematemesis; (3) perforation diagnosed by visualization of contrast medium or air around the margins of the bile duct and duodenum by fluoroscopy or computed tomography; (4) acute cholecystitis diagnosed according to the Tokyo guidelines of cholecystitis17; (5) liver abscess diagnosed based on the presence of a marked inflammatory reaction and detection of a new hepatic mass by imaging examinations; and (6) other AEs, which were defined as any deviation from the expected clinical course after SEMS placement believed to be related to the ES procedure. The severity of post–SEMS placement pancreatitis was graded using a modified Cotton et al16 system in which the number of hospitalization days was replaced by the number of fasting days. Bleeding and perforation were graded according to the Cotton et al16 grading system, and cholecystitis and liver abscess were graded according to the Cotton et al16 grading system for infection (cholangitis).16 The time to SEMS dysfunction was calculated from the date of SEMS placement to the date of recurrent cholangitis diagnosis based on the Tokyo guidelines of cholangitis.18 Patients who did not develop cholangitis before the end of the follow-up period, died, or underwent SEMS removal for any reason other than SEMS occlusion were censored. Patient survival time was defined as the interval between SEMS placement and patient death, and patients alive at the end of the followup period were censored.
Outcome Measures and Statistical Analysis Our chief concern was whether ES was necessary to prevent post–SEMS placement pancreatitis because otherwise it would be desirable to skip this procedure to decrease the risk of bleeding and/or perforation and to reduce the cost and procedure time. We hypothesized that SEMS placement without ES would be noninferior to SEMS placement with ES with respect to the rates of early AEs in patients with unresectable pancreatic cancer. The primary outcome measure was the frequency of early AEs specifically related to the presence or absence of ES (pancreatitis, bleeding, and perforation). The ontreatment analysis included data from patients who underwent randomization and successful SEMS placement. The difference in the proportion of patients with an early AE specifically related to the presence or absence of ES between the non-ES and ES groups and corresponding 95% confidence interval (CI) were estimated. Noninferiority of non-ES was defined as an upper limit of the
1154 Hayashi et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
95% CI that was smaller than the prespecified noninferiority margin of 10%. The previously reported rates of early SEMS placement-related AEs varied widely (0%–25%).2–5,6–15 However, during the planning of the study protocol (August 2009), only 2 prospective studies concerning covered SEMS placement for the treatment of distal malignant biliary stricture were available.4,12 In these studies, the rates of AEs specifically related to the presence or absence of ES ranged from 4% to 8.8% (mean, 6.4%). Assuming an AE rate of 6% in both the non-ES and ES groups, we calculated that 86 patients per group were required to establish noninferiority with a noninferiority margin of 10%, 80% power, and a 2-sided significance level of 5%. Anticipating potential study drop-outs, we enrolled 100 patients per group (200 patients total). The secondary outcome measures were as follows: endoscopic procedure time, all early and late SEMSrelated AEs, change in serum amylase level 24 hours after SEMS placement, time to SEMS dysfunction, cause of SEMS dysfunction, and patient survival. To further examine the potential benefits and disadvantages of SEMS placement without ES, the superiority of non-ES vs ES in each outcome measure was analyzed in an intention-to-treat (ITT) population. The data are presented as the number and percentage, mean and standard deviation, or median and range as appropriate. Categoric variables were compared using
the Fisher exact test. Continuous variables with normal and skewed data distributions were compared using the Student t test and the Mann–Whitney U test, respectively. The time to SEMS dysfunction and patient survival were estimated using the Kaplan–Meier method and compared using the log-rank test. Serum amylase level was expressed as the geometric mean, and 95% CI and was compared between the non-ES and ES groups using a 2-way analysis of variance. All statistical analyses were performed using SPSS (version 19.0; IBM, Tokyo, Japan) and SAS (version 9.3; SAS, Tokyo, Japan) software. For all analyses, a P value less than .05 was considered statistically significant.
Results Patient Demographics and Self-Expandable Metallic Stent Placement–Related Data Between August 2010 and May 2012, there were 519 patients with biliary stricture caused by pancreatic cancer who were assessed for eligibility, and 319 patients were excluded. The 200 included patients were allocated randomly into equal-sized non-ES and ES groups. Biliary cannulation failure precluded SEMS placement in 2 and 4 patients in the non-ES and ES groups, respectively (Figure 1). Patient characteristics and SEMS
Figure 1. Flow diagram of patient enrollment, randomization, and follow-up evaluation.
June 2015
Sphincterotomy Before Stent Placement 1155
Table 2. Demographic and SEMS Placement-Related Data
Patient characteristic Age, mean SD, y Sex, male/female, n Initial drainage, n Endoscopic drainage (EBS/ENBD), n Percutaneous drainage (PTBD/PTGBD), n History of acute pancreatitis, n Gallbladder present, n Serum amylase level before the procedure, geometric mean (95% CI), IU/L MPD obstruction, n Maximum diameter of the proximal MPD, mean SD, mm SEMS placement–related data ERCP experience of endoscopists, median (range), y Type of duodenoscope (JF260V/TJF260V/TJF240/JF240/JF200; Olympus Medical Systems Corp, Tokyo, Japan), n Use of medical prophylaxis against post–SEMS placement pancreatitis (none/ulinastatin/gabexate mesilate/nafamostat mesilate/nonsteroidal anti-inflammatory drugs), n Guidewire inserted into the pancreatic duct, n Contrast medium injected into the pancreatic duct, n Extent of opacified pancreatic duct (main/branch/acinar), n Length of the biliary stricture, mean SD, mm Use of the rendezvous method, n SEMS length, 40/60/80/failure, mm, n Procedure time, mean SD, s
Non-ES (n ¼ 100)
ES (n ¼ 100)
P value
70.54 10.4 49/51 68 50/16 2/0 3 93 55.27 (48.19–63.38) 83 6.55 3.41
72.2 10.6 53/47 70 53/13 3/1 7 91 61.71 (53.21–71.56) 80 6.91 4.82
.27 .67 .88
9 (3–31) 52/43/2/0/3
9 (1–31) 51/45/1/1/2
.90 .92
28/44/17/11/0
31/45/17/7/0
.81
18 11 8/3/0 26.14 12.13 1 0/90/8/2 380.1 208.5
16 9 5/4/0 27.00 14.03 2 1/82/13/4 553.7 324.8
.85 .81
.21 .80 .28 .58 .54
.64 .62 .36 <.001
EBS, endoscopic biliary stenting; ENBD, endoscopic nasobiliary drainage; MPD, main pancreatic duct; PTBD, percutaneous transhepatic biliary drainage; PTGBD, percutaneous transhepatic gallbladder drainage.
placement–related data are summarized in Table 2. Patient characteristics were not significantly different between the non-ES and ES groups except for procedure time. Predictably, procedure time was significantly shorter in the non-ES group than in the ES group (387.9 203.3 vs 576.7 310.3 s; P < .001).
Rates of Early Adverse Events Specifically Related to the Presence or Absence of Endoscopic Sphincterotomy In the non-ES group, early AEs specifically related to the presence or absence of ES were observed in Table 3. Rates of Early AEs Specifically Related to the Presence or Absence of ES
Early AEs, n (%) Pancreatitis, n Mild/moderate/ severe, n Moderate bleeding, n Mild perforation, n
Non-ES (n ¼ 98)
ES (n ¼ 96)
9 (9.2) 8 3/3/2
10 (10.4) 9 6/3/0
0
1
1
0
Difference between non-ES and ES (95% CI) -1.2% (-9.6 to 7.1)
NOTE. Boldface indicates primary outcome measure of the study.
9 patients including post–SEMS placement pancreatitis in 8 patients and mild perforation during guidewire manipulation in 1 patient (Table 3). In the ES group, early AEs specifically related to the presence or absence of ES occurred in 10 patients, including post–SEMS placement pancreatitis in 9 patients and moderate bleeding in 1 patient (Table 3). Among the patients in whom SEMS placement was successful (on-treatment population), the percentage point difference in the rates of early AEs specifically related to the presence or absence of ES between the non-ES and ES groups was -1.2% (95% CI, 9.6% to 7.1%), indicating that non-ES was noninferior to ES. Furthermore, the rates for early and late SEMS placement–related AEs were not significantly different between the 2 groups in the ITT population (Supplementary Tables 1–3).
Change in Serum Amylase Level After Self-Expandable Metallic Stent Placement In the non-ES group, the serum amylase level increased significantly after SEMS placement from 54.91 IU/L (95% CI, 47.80–63.09 IU/L) to 116.97 IU/L (95% CI, 92.84–147.38 IU/L; P < .001). The serum amylase level also increased significantly after SEMS placement in the ES group from 63.14 IU/L (95% CI, 54.21–73.5 IU/L) to 125.76 IU/L (95% CI, 93.53–161.56 IU/L; P < .001). However, 2-way analysis of variance showed no significant difference in the change in serum amylase between
1156 Hayashi et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
Discussion
Figure 2. Change in serum amylase level after SEMS placement in the non-ES and ES groups. The large black circles represent the geometric mean of the serum amylase level, and the error bars indicate the associated 95% confidence interval.
the 2 groups in the ITT population (interaction P ¼ .38) (Figure 2).
Time to Self-Expandable Metallic Stent Dysfunction and Patient Survival Time The last patient was enrolled in May 2012, and the follow-up period ended in November 2012. The median follow-up duration was 192 days (range, 14–810 d) and 219.5 days (range, 14–757 d) in the non-ES and ES groups, respectively. The median time to SEMS dysfunction was not significantly different between the non-ES and ES groups in the ITT population (>594 vs 541 d; P ¼ .88). The median patient survival time also was not significantly different between the non-ES and ES groups in the ITT population (202 vs 255 d, respectively; P ¼ .20) (Supplementary Table 4, Figure 3).
We performed a noninferiority study to determine whether omitting ES before SEMS placement for distal biliary stricture caused by unresectable pancreatic cancer would increase the frequency of AEs specifically related to the presence or absence of ES (pancreatitis, bleeding, and perforation) and found that ES omission did not affect the rates of AEs specifically related to the presence or absence of ES. Thus, the present study overturns the conventional wisdom that ES prevents pancreatitis after SEMS placement. Generally, mechanical injury from manipulation of the papilla of Vater or the pancreatic duct is considered the major cause of post-ERCP pancreatitis. In addition to careful manipulation and adequate skill level of the endoscopist, nonsteroidal anti-inflammatory drugs and pancreatic stenting also can prevent pancreatitis. Pancreatic juice congestion caused by pancreatic orifice occlusion is considered another etiologic cause of postSEMS placement pancreatitis. ES theoretically could prevent such congestion by decreasing the pressure that the SEMS exerts toward the pancreatic orifice. However, the increase in serum amylase level, a common measure of pancreatic juice congestion, was similar between the non-ES and ES groups, indicating that ES does not protect the pancreatic duct orifice after SEMS placement. There are 2 possible explanations for the lack of protective effect: the expansive force of the SEMS may be so strong that ES cannot protect the pancreatic orifice, and/ or the axial (straightening) force of the SEMS may directly affect the pancreatic orifice rather than being limited to the incised duodenal mucosa as previously reported.19,20 The rates of perforation and bleeding in the ES group were only 0% and 1%, respectively, and were comparable with previously reported rates for
Figure 3. (A) Kaplan– Meier graph showing the time to SEMS dysfunction. (B) Kaplan–Meier graph showing patient survival.
June 2015
covered SEMS (0%–0.7% and 0%–2%, respectively).4,5,7–9,11–15 These rates are relatively low compared with those for ES for extraction of bile duct stones, and this may be attributable to the wrapping and tamponade effect of covered SEMS because the covering membrane may seal tiny perforations and other minor vessel injuries that occur during ES. A significant limitation of the present study was the unmasked assessment. It was impossible to mask the physicians (assessors) who evaluated the AEs to the treatment allocation because of the shortage of staff, especially in some of the small, rural, secondary referral hospitals. In general, failure to mask studies investigating the superiority of a newly developed procedure is thought to impose a high risk of substantial assessor bias. However, the purpose of this study was to show the noninferiority of non-ES relative to ES before SEMS placement. ES was a frequently performed procedure at the time the study was conducted; therefore, we would have expected any assessor bias to be slanted toward its presumed benefits. Therefore, we doubt that the unmasked assessment substantially affected our results. Another limitation was the undefined administration of prophylactics for pancreatitis. We cannot completely rule out the potential influence of prophylactic use on the pancreatitis rate. However, we believe that non-ES is noninferior to ES despite the use of prophylactics because the preventive effect of the prophylactics used in the present study has not been proven in any clinical study and the proportion of patients receiving the different types of prophylactics was not significantly different between the non-ES and ES groups. Our study also was not multinational. In prospective multicenter clinical studies in Japan, the rate of post-ERCP pancreatitis in high-risk and unselected subjects has been reported to be 11.3% and 3.5%, respectively.21,22 These rates are similar to those reported in other ethnic groups.23,24 In addition, one of the major causes of post–metallic stent placement pancreatitis is the occlusion of the pancreatic orifice. The occurrence of pancreatic orifice occlusion is not influenced by the endoscopist’s skill and may not be affected by patient race or ethnicity. We believe our study findings are applicable to all patients regardless of ethnicity, and thus, will impact clinical practice worldwide. In conclusion, the present study showed that ES does not affect the incidence of AEs, SEMS patency rate, or patient survival. Therefore, we conclude that ES is not necessary before SEMS placement for distal biliary stricture caused by unresectable pancreatic cancer.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2015.01.008.
Sphincterotomy Before Stent Placement 1157
References 1. Porta M, Fabregat X, Malats N, et al. Exocrine pancreatic cancer: symptoms at presentation and their relation to tumour site and stage. Clin Transl Oncol 2005;7:189–197. 2. Davids PH, Groen AK, Rauws EA, et al. Randomised trial of selfexpanding metal stents versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:1488–1492. 3. Knyrim K, Wagner HJ, Pausch J, et al. A prospective, randomized, controlled trial of metal stents for malignant obstruction of the common bile duct. Endoscopy 1993;25:207–212. 4. Soderlund C, Linder S. Covered metal versus plastic stents for malignant common bile duct stenosis: a prospective, randomized, controlled trial. Gastrointest Endosc 2006;63:986–995. 5. Isayama H, Yasuda I, Ryozawa S, et al. Results of a Japanese multicenter, randomized trial of endoscopic stenting for nonresectable pancreatic head cancer (JM-test): Covered Wallstent versus DoubleLayer stent. Dig Endosc 2011;23:310–315. 6. Artifon ELA, Sakai P, Ishioka S, et al. Endoscopic sphincterotomy before deployment of covered metal stent is associated with greater complication rate: a prospective randomized control trial. J Clin Gastroenterol 2008;42:815–819. 7. Isayama H, Kawabe T, Nakai Y, et al. Management of distal malignant biliary obstruction with the ComVi stent, a new covered metallic stent. Surg Endosc 2010;24:131–137. 8. Kullman E, Frozanpor F, Söderlund C, et al. Covered versus uncovered self-expandable nitinol stents in the palliative treatment of malignant distal biliary obstruction: results from a randomized, multicenter study. Gastrointest Endosc 2010;72:915–923. 9. Tsuchiya T, Itoi T, Gotoda T, et al. A multicenter prospective study of the short-term outcome of a newly developed partially covered self-expandable metallic biliary stent (WallFlex(®)). Dig Dis Sci 2011;56:1889–1895. 10. Aadam AA, Evans DB, Khan A, et al. Efficacy and safety of selfexpandable metal stents for biliary decompression in patients receiving neoadjuvant therapy for pancreatic cancer: a prospective study. Gastrointest Endosc 2012;76:67–75. 11. Isayama H, Mukai T, Itoi T, et al. Comparison of partially covered nitinol stents with partially covered stainless stents as a historical control in a multicenter study of distal malignant biliary obstruction: the WATCH study. Gastrointest Endosc 2012; 76:84–92. 12. Isayama H, Komatsu Y, Tsujino T, et al. A prospective randomized study of “covered” versus “uncovered” diamond stents for the management of distal malignant biliary obstruction. Gut 2004;53:729–734. 13. Costamagna G, Tringali A, Reddy DN, et al. A new partially covered nitinol stent for palliative treatment of malignant bile duct obstruction: a multicenter single-arm prospective study. Endoscopy 2011;43:317–324. 14. Telford JJ, Carr-Locke DL, Baron TH, et al. A randomized trial comparing uncovered and partially covered self-expandable metal stents in the palliation of distal malignant biliary obstruction. Gastrointest Endosc 2010;72:907–914. 15. Ung KA, Stotzer PO, Nilsson A, et al. Covered and uncovered self-expandable metallic Hanarostents are equally efficacious in the drainage of extrahepatic malignant strictures. Results of a double-blind randomized study. Scand J Gastroenterol 2013; 48:459–465. 16. Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37:383–393.
1158 Hayashi et al 17. Hirota M, Takada T, Kawarada Y, et al. Diagnostic criteria and severity assessment of acute cholecystitis: Tokyo guidelines. J Hepatobiliary Pancreat Surg 2007;14:78–82. 18. Wada K, Takada T, Kawarada Y, et al. Diagnostic criteria and severity assessment of acute cholangitis: Tokyo guidelines. J Hepatobiliary Pancreat Surg 2007;14:52–58. 19. Kawakubo K, Isayama H, Nakai Y, et al. Risk factors for pancreatitis following transpapillary self-expandable metal stent placement. Surg Endosc 2012;26:771–776. 20. Tarnasky PR, Cunningham JT, Hawes RH, et al. Transpapillary stenting of proximal biliary strictures: does biliary sphincterotomy reduce the risk of postprocedure pancreatitis? Gastrointest Endosc 1997;45:46–51. 21. Sofuni A, Maguchi H, Itoi T, et al. Prophylaxis of postendoscopic retrograde cholangiopancreatography pancreatitis by an endoscopic pancreatic spontaneous dislodgement stent. Clin Gastroenterol Hepatol 2007;5:1339–1346. 22. Kawakami H, Maguchi H, Mukai T, et al. A multicenter, prospective, randomized study of selective bile duct cannulation performed by multiple endoscopists: the BIDMEN study. Gastrointest Endosc 2012;75:362–372. 23. ASGE Standards of Practice CommitteeAnderson Michelle A, Fisher Laurel, et al. Complication of ERCP. Gastrointest Endosc 2012;75:467–473. 24. Dumonceau JM, Andriulli A, Elmunzer BJ, et al. Prophylaxis of post-ERCP pancreatitis: European Society of Gastrointestinal Endoscopy (ESGE) guideline - updated June 2014. Endoscopy 2014;46:799–815.
Reprint requests Address requests for reprints to: Tsuyoshi Hayashi, MD, Department of Medical Oncology and Hematology, Sapporo Medical University School of Medicine, South-1, West-16, Chuo-ku, Sapporo 060-8543, Hokkaido, Japan. e-mail: [email protected]; fax: (81) 11-612-7987.
Clinical Gastroenterology and Hepatology Vol. 13, No. 6 Acknowledgments The authors gratefully acknowledge the assistance of Michihiro Ono, MD, Makoto Yoshida, MD, Fumito Tamura, MD, Shingo Tanaka, MD, Hiroto Horiguchi, MD, and Junji Kato, MD (Department of Medical Oncology and Hematology, Sapporo Medical University); Masaki Kuwatani, MD, Kazumichi Kawakubo, MD, Kazunori Eto, MD, Shin Haba, MD, Taiki Kudo, MD, Yoko Abe, MD, Shuhei Kawahata, MD, Masahiro Asaka, MD, and Naoya Sakamoto, MD (Department of Gastroenterology and Hepatology, Hokkaido University Graduate School of Medicine); Hiroyuki Maguchi, MD, Kuniyuki Takahashi, MD, Akio Katanuma, MD, Kei Yane, MD, Shin Kato, MD, Maki Kaneko, MD, Shunpei Hashigo, MD, Ryo Harada, MD, and Ryusuke Kato, MD (Center for Gastroenterology, Teine-Keijinkai Hospital); Hiroaki Yamato, MD (Department of Gastroenterology, Hakodate Municipal Hospital); Hiroyuki Isshiki, MD, Kentaro Kawakami, MD, and Takafumi Naito, MD (Department of Gastroenterology, Muroran City General Hospital); Toru Kawamoto, MD, Junpei Sasajima, MD, and Yoshiaki Sugiyama, MD (Division of Gastroenterology and Hematology/ Oncology, Department of Medicine, Asahikawa Medical University); Naoki Uemura, MD (Department of Gastroenterology, Steel Memorial Muroran Hospital); Hiroyuki Miyakawa, MD (Department of Bilio-Pancreatology, Sapporo Kosei General Hospital); Takahiro Osuga, MD (Department of Gastroenterology, National Hospital Organisation Hokkaido Cancer Center); Ryusuke Matsumoto, MD (Third Department of Internal Medicine, Obihiro-Kosei General Hospital); Atsushi Chiba, MD (Department of Internal Medicine, Asahikawa City Hospital); Masahiro Shitani, MD (Department of Gastroenterology, Rheumatology, and Clinical Immunology, Sapporo Medical University); Keisuke Shinada, MD, and Akihiko Yumino, MD (Department of Gastroenterology, Keiwakai Ebetsu Hospital); Keisuke Yamakita, MD, and Masako Imazawa, MD (Division of Metabolism and Biosystemic Science, Department of Internal Medicine, Asahikawa Medical University); Manabu Onodera, MD (Department of Gastroenterology, Abashiri-Kosei General Hospital); Sho Takahashi, MD (Department of Gastroenterology, Otaru Kyokai Hospital); Tetsuya Miseki, MD (Department of Internal Medicine, Obihiro Daiichi Hospital); Minoru Uebayashi, MD (Department of Gastroenterology, Kitami Red Cross Hospital); Hideyuki Ihara, MD (The Center for Digestive Diseases, Tonan Hospital); Takuro Machida, MD, and Shuji Ikari, MD (Department of Internal Medicine, Hokkaido Gastroenterology Hospital); Nobuaki Akakura, MD (Department of Gastroenterology, Sapporo Medical Center Nippon Telegraph and Telephone East Corporation); Takehiro Kukitsu, MD (Department of Gastroenterology, Rumoi Municipal Hospital); and Kunihiro Takanashi, MD (Department of Gastroenterology, Otaru Ekisaikai Hospital). Conflicts of interest The authors disclose no conflicts. Funding Tsuyoshi Hayashi is supported in part by a grant from The Japanese Foundation for Research and Promotion of Endoscopy.
June 2015
Sphincterotomy Before Stent Placement 1158.e1
Supplementary Table 1. Rates of Early AEs Related to SEMS Placement
Early AEs, n (%) Pancreatitis, n Mild/moderate/severe, n Moderate bleeding, n Mild perforation, n Moderate liver abscess, n Epigastric pain, n Vomiting, n
Non-ES (n ¼ 100)
ES (n ¼ 100)
15 (15) 8 3/3/2 0 1 3 3 0
15 (15) 9 6/3/0 1 0 1 2 2
P value 1.0
Supplementary Table 2. Rates of Late AEs Related to SEMS Placement
Late AEs, n Severe pancreatitis, n Bleeding, n Mild/moderate/severe, n Moderate liver abscess, n Duodenal ulcer, n
Non-ES (n ¼ 100)
ES (n ¼ 100)
5 0 1 0/1/0 1 1
6 1 1 0/0/1 0 0
Supplementary Table 3. Rates of Early and Late-Onset Cholecystitis
P value 1.0
Early onset, n (%) Mild/moderate/severe, n Late onset, n (%) Mild/moderate/severe, n
Non-ES (n ¼ 93)
ES (n ¼ 91)
4 (4.3) 0/3/1 4 (4.3) 0/4/0
1 (1.1) 0/1/0 1 (1.1) 0/1/0
P value .37 .37
1158.e2 Hayashi et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
Supplementary Table 4. Time to SEMS Dysfunction and Patient Survival Time
Time to SEMS dysfunction, median (95% CI), da Frequency of cholangitis, n (%) Cause of cholangitis, n (%) Food impaction Sludge formation Intestinal obstruction Proximal migration Distal migration Tumor ingrowth Tumor overgrowth Bleeding Unknown Patient survival, median (95% CI), db
10 5 2 2 1 1 3 0 1 202
Non-ES
ES
P value
>594 25
541 (353.4–728.7) 25
.88
(40) (20) (8) (8) (4) (4) (12) (4) (170.5–233.5)
10 4 3 3 1 2 0 1 1 255
(40) (16) (12) (12) (4) (8) (4) (4) (187.9–322.1)
.20
NOTE. In the non-ES group, the SEMS was removed in 7 patients. Fifty-two and 14 patients were censored at the time of patient death or at the end of the followup period, respectively. In the ES group, the SEMS was removed in 2 patients. Forty-nine and 20 patients were censored at the time of patient death or the end of the follow-up period, respectively. SEMS dysfunction developed in 25 patients in each group. The median time to SEMS dysfunction could not be calculated because it was longer than 594 days in the non-ES group and was 541 days in the ES group as estimated by the Kaplan–Meier method; this difference was not significant (P ¼ .88). The main causes of cholangitis were food impaction and sludge formation. The frequencies of all causes of SEMS dysfunction were similar between the 2 groups. Death from pancreatic cancer progression occurred in 78 patients in the non-ES and in 67 in the ES group. The median patient survival time was 202 days in the non-ES group and 255 days in the ES group, a nonsignificant difference (P ¼ .20). a The median (95% CI) time to SEMS dysfunction was estimated by the Kaplan–Meier method. b The median (95% CI) patient survival time was estimated by the Kaplan–Meier method.
PANCREAS, BILIARY TRACT, AND LIVER No Benefit of Endoscopic Sphincterotomy Before Biliary Placement of Self-Expandable Metal Stents for Unresectable Pancreatic Cancer Tsuyoshi Hayashi,* Hiroshi Kawakami,‡ Manabu Osanai,§ Hirotoshi Ishiwatari,* Hirohito Naruse,k Hiroyuki Hisai,¶ Nobuyuki Yanagawa,# Hiroyuki Kaneto,** Kazuya Koizumi,‡‡ Tamaki Sakurai,§§ and Tomoko Sonodakk *Department of Medical Oncology and Hematology, kkDepartment of Public Health, Sapporo Medical University School of Medicine, Sapporo, Japan; ‡Department of Gastroenterology and Hepatology, Hokkaido University Graduate School of Medicine, Sapporo, Japan; §Center for Gastroenterology, Teine-Keijinkai Hospital, Sapporo, Japan; kDepartment of Gastroenterology, Hakodate Municipal Hospital, Hakodate, Japan; ¶Department of Gastroenterology, Japan Red Cross Date General Hospital, Date, Japan; #Department of Gastroenterology, Asahikawa-Kosei General Hospital, Asahikawa, Japan; **Department of Gastroenterology, Muroran City General Hospital, Muroran, Japan; ‡‡Division of Gastroenterology and Hematology/Oncology, Department of Medicine, Asahikawa Medical University, Asahikawa, Japan; §§Department of Gastroenterology, Steel Memorial Muroran Hospital, Muroran, Japan BACKGROUND & AIMS:
Endoscopic sphincterotomy (ES) is performed routinely before self-expandable metallic stents (SEMS) are placed in malignant distal biliary strictures to prevent postprocedural pancreatitis. However, it is not clear whether ES actually prevents pancreatitis or affects other adverse events (AEs). We conducted a noninferiority trial to examine the necessity of ES before SEMS placement.
METHODS:
Two hundred patients with distal biliary strictures caused by unresectable pancreatic cancer were assigned randomly to groups that received ES or did not receive ES (non-ES) before SEMS placement, at 25 hospitals in Hokkaido, Japan, from August 2010 through November 2012. The primary outcome was early AEs (£30 d) specifically related to the presence or absence of ES (pancreatitis, bleeding, or perforation). Secondary outcomes measured included the effect of ES omission on time to SEMS dysfunction and patient survival times.
RESULTS:
The proportions of patients with early AEs were 9.2% in the non-ES group and 10.4% in the ES group (a difference of 1.2%, noninferior). The median times to SEMS dysfunction was longer than 594 days in the non-ES group and 541 days in the ES group (P [ .88). The median overall survival times were 202 in the ES group vs 255 days in the non-ES group; P [ .20).
CONCLUSIONS:
ES before SEMS does not affect the incidence of AEs, SEMS patency, or patient survival times. Our data provide no evidence for a benefit of ES to patients undergoing SEMS placement for a biliary stricture caused by pancreatic cancer. UMIN clinical trials registry number: 000004044.
Keywords: Endoscopic Sphincterotomy; Self-Expandable Metallic Stent; Adverse Event; Pancreas; Distal Biliary Stricture.
See editorial on page 1159. istal biliary stricture is a common comorbidity of unresectable pancreatic cancer. Because resection is not feasible at the time of diagnosis for one third of pancreatic cancers,1 palliation of biliary stricture is important to improve quality of life and allow initiation of chemotherapy. Decompression of biliary strictures usually is accomplished by endoscopic stent placement. Self-expandable metallic stent (SEMS) placement is now a standard procedure for distal biliary stricture caused
D
by unresectable malignant lesions and achieves longer stent patency than plastic stent placement because of the larger diameter of the SEMS.2–5 However, pancreatitis is a common postprocedural complication. Many endoscopists believe that this adverse event (AE) can Abbreviations used in this paper: AE, adverse event; CI, confidence interval; ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic sphincterotomy; ITT, intention-to-treat; SEMS, self-expandable metallic stent. © 2015 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2015.01.008
1152 Hayashi et al
be prevented by endoscopic sphincterotomy (ES). To date, only 1 prospective randomized controlled trial has examined the influence of ES on the rate of AEs associated with SEMS placement for malignant biliary stricture. They concluded that ES is disadvantageous in patients with malignant biliary stricture based on the lower rate of SEMS-related AEs in the non-ES group compared with the ES group (11% vs 49%; P ¼ .006).6 However, the combined rate of bleeding and perforation in the ES group was 24.3%,6 which was exceedingly high relative to other reports of SEMS placement with ES (0%–1%).2–5,7–11 Furthermore, post–SEMS placement pancreatitis was not observed in either the non-ES or ES groups.6 This is in contrast to several previous studies that have reported that pancreatitis is the most frequent early post-SEMS placement AE (0%–8.8%).2–5,7–15 Because these results do not reflect the typical clinical practice experience and impaired pancreatic outflow is considered a major cause of pancreatitis, the majority of endoscopists continue to perform ES routinely before SEMS placement to minimize the influence of the strong expansive force of the SEMS on the pancreatic duct orifice. In Japan, some endoscopists believe ES before SEMS placement is useless, especially in pancreatic cancer patients in whom exocrine function already has ceased because of a main pancreatic duct obstruction. Therefore, we conducted a multicenter, prospective, randomized, controlled trial to investigate the influence of ES on the incidence of AEs associated with SEMS placement for biliary stricture in patients with unresectable pancreatic cancer (the multicenter, prospective, randomized, controlled trial to investigate the influence of endoscopic sphincterotomy on the incidence of adverse events associated with self-expandable metallic stent placement for biliary stricture in patients with unresectable pancreatic cancer trial in Hokkaido).
Patients and Methods Study Design Twenty-five hospitals in Hokkaido, Japan, participated in the study including 3 educational, 5 tertiary referral, and 17 secondary referral hospitals. The treatment (non-ES or ES) was assigned randomly upon enrollment by a 1:1 centralized allocation using the minimization method. The allocation and data collection were performed using a web-based system that was unavailable to the investigators. Randomization was concealed from patients, endoscopists, and physicians. However, patients, endoscopists, and physicians were unmasked to the treatment allocation. The study protocol was approved by the Institutional Review Board of each participating hospital and registered with the University Hospital Medical Information Network Clinical Trials Registry (http://www.umin.ac.jp/ctr/index/htm/; no. University Hospital Medical Information Network
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
000004044). All authors had access to the study data and reviewed and approved the final manuscript.
Patients Complete medical history, physical examination, Eastern Cooperative Oncology Group performance status, blood examination, contrast enhanced-computed tomography and/or magnetic resonance imaging study findings, endoscopic examinations, and cytohistologic diagnosis were used to assess the eligibility of patients with distal biliary stricture caused by unresectable pancreatic cancer. Patients meeting all of the inclusion and none of the exclusion criteria were considered eligible (Table 1). Written informed consent to participate in the study was obtained from all participants at each hospital before randomization.
Self-Expandable Metallic Stent Placement SEMS placement was performed as an inpatient procedure in the participating hospitals, each of which had at least 1 experienced endoscopist (defined as 10 years of postresidency experience, 2000 endoscopic retrograde cholangiopancreatographies (ERCPs), or an annual ERCP caseload >200). All units permitted trainees to perform SEMS placement under the supervision of an experienced endoscopist. All patients Table 1. Eligibility Criteria Inclusion criteria Distal biliary stricture resulting from unresectable pancreatic cancer diagnosed based on contrast-enhanced computed tomography and/or magnetic resonance imaging findings Histologically or cytologically confirmed pancreatic cancer Age 20 y Ability to understand and willingness to sign a written statement of informed consent Exclusion criteria Moderate or severe cholangitis according to the Tokyo guidelines of cholangitis18 Involvement of the ampulla of Vater confirmed by endoscopy Billroth II or Roux-en-Y reconstruction after gastrectomy Intestinal obstruction toward the anal direction of the ampulla of Vater Pancreaticobiliary maljunction Biliary stricture extending to the hilum History of placement of a transpapillary biliary drainage tube with a bore larger than 8F History of previous endoscopic sphincterotomy or precut sphincterotomy Current use of anticoagulant or antiplatelet drugs Prothrombin time–international normalized ratio 1.5 and/or 50,000 platelets/mL ECOG performance status of 4 Severe heart or pulmonary disease Any disorder that compromised the patient’s ability to provide written informed consent and/or comply with the study procedure
ECOG, Eastern Cooperative Oncology Group.
June 2015
underwent SEMS placement under sedation using a combination of diazepam, midazolam, and/or pethidine hydrochloride. Various types of ERCP catheters compatible with the 0.035-inch hydrophilic guidewire were used. The rendezvous method of biliary cannulation was permitted in patients who underwent percutaneous biliary drainage; however, precut sphincterotomy was absolutely prohibited, even for difficult biliary cannulations, in both the non-ES and ES groups. Pancreatic sphincterotomy and prophylactic pancreatic stent placement also were prohibited. In the ES group, after successful guidewire placement and cholangiography, an incision was made using a standard traction-type sphincterotome with a blended current and was continued until visual recognition of bile outflow or before reaching the upper edge of the oral protrusion in cases without outflow. In both groups, a 10-mm diameter WallFlex Biliary RX Partially Covered Stent (Boston Scientific Japan, Tokyo, Japan) was inserted over the guidewire under fluoroscopic guidance. A 40-, 60-, or 80-mm stent was chosen according to the location and length of the biliary stricture as estimated by withdrawing the catheter or guidewire from the proximal end of the stricture to the ampulla of Vater. The SEMS routinely was placed across the papilla with approximately 10 mm of its distal end exposed to the duodenal lumen.
Follow-Up Evaluation After the procedure, all patients fasted until the attending physician confirmed that no AEs had occurred. A blood examination routinely was performed 24 hours after the procedure. Patients who experienced AEs were treated in the hospital as necessary. After discharge, blood parameters, clinical signs, and symptoms of the patients were monitored by a physician at an outpatient clinic at least monthly until patient death or the end of the study period, and all AEs, cholangitis incidence, and patient survival were recorded by the participating hospitals. All clinical information was collected consistently, even for patients who visited or were transferred to affiliated hospitals. Patients who developed recurrent cholangitis were re-admitted to the participating hospital, and endoscopy was performed to investigate the cause and for palliative treatment. The follow-up period ended 6 months after enrollment of the last patient.
Definitions The endoscopic procedure time was defined as the time from guidewire placement to confirmation of successful SEMS placement by visual recognition of bile outflow from the distal end of the SEMS by duodenoscopy or disappearance of the contrast medium from the common bile duct by fluoroscopy. Early and late AEs were defined as any SEMS placement–related AE
Sphincterotomy Before Stent Placement 1153
occurring within 30 days or more than 31 days after SEMS placement, respectively. AEs were defined as follows: (1) post–SEMS placement pancreatitis diagnosed according to the Cotton et al16 criteria (ie, new or worsening upper abdominal pain lasting >24 hours after the procedure in conjunction with hyperamylasemia >3 times the upper limit of the normal range); (2) bleeding diagnosed by the presence of melena or hematemesis; (3) perforation diagnosed by visualization of contrast medium or air around the margins of the bile duct and duodenum by fluoroscopy or computed tomography; (4) acute cholecystitis diagnosed according to the Tokyo guidelines of cholecystitis17; (5) liver abscess diagnosed based on the presence of a marked inflammatory reaction and detection of a new hepatic mass by imaging examinations; and (6) other AEs, which were defined as any deviation from the expected clinical course after SEMS placement believed to be related to the ES procedure. The severity of post–SEMS placement pancreatitis was graded using a modified Cotton et al16 system in which the number of hospitalization days was replaced by the number of fasting days. Bleeding and perforation were graded according to the Cotton et al16 grading system, and cholecystitis and liver abscess were graded according to the Cotton et al16 grading system for infection (cholangitis).16 The time to SEMS dysfunction was calculated from the date of SEMS placement to the date of recurrent cholangitis diagnosis based on the Tokyo guidelines of cholangitis.18 Patients who did not develop cholangitis before the end of the follow-up period, died, or underwent SEMS removal for any reason other than SEMS occlusion were censored. Patient survival time was defined as the interval between SEMS placement and patient death, and patients alive at the end of the followup period were censored.
Outcome Measures and Statistical Analysis Our chief concern was whether ES was necessary to prevent post–SEMS placement pancreatitis because otherwise it would be desirable to skip this procedure to decrease the risk of bleeding and/or perforation and to reduce the cost and procedure time. We hypothesized that SEMS placement without ES would be noninferior to SEMS placement with ES with respect to the rates of early AEs in patients with unresectable pancreatic cancer. The primary outcome measure was the frequency of early AEs specifically related to the presence or absence of ES (pancreatitis, bleeding, and perforation). The ontreatment analysis included data from patients who underwent randomization and successful SEMS placement. The difference in the proportion of patients with an early AE specifically related to the presence or absence of ES between the non-ES and ES groups and corresponding 95% confidence interval (CI) were estimated. Noninferiority of non-ES was defined as an upper limit of the
1154 Hayashi et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
95% CI that was smaller than the prespecified noninferiority margin of 10%. The previously reported rates of early SEMS placement-related AEs varied widely (0%–25%).2–5,6–15 However, during the planning of the study protocol (August 2009), only 2 prospective studies concerning covered SEMS placement for the treatment of distal malignant biliary stricture were available.4,12 In these studies, the rates of AEs specifically related to the presence or absence of ES ranged from 4% to 8.8% (mean, 6.4%). Assuming an AE rate of 6% in both the non-ES and ES groups, we calculated that 86 patients per group were required to establish noninferiority with a noninferiority margin of 10%, 80% power, and a 2-sided significance level of 5%. Anticipating potential study drop-outs, we enrolled 100 patients per group (200 patients total). The secondary outcome measures were as follows: endoscopic procedure time, all early and late SEMSrelated AEs, change in serum amylase level 24 hours after SEMS placement, time to SEMS dysfunction, cause of SEMS dysfunction, and patient survival. To further examine the potential benefits and disadvantages of SEMS placement without ES, the superiority of non-ES vs ES in each outcome measure was analyzed in an intention-to-treat (ITT) population. The data are presented as the number and percentage, mean and standard deviation, or median and range as appropriate. Categoric variables were compared using
the Fisher exact test. Continuous variables with normal and skewed data distributions were compared using the Student t test and the Mann–Whitney U test, respectively. The time to SEMS dysfunction and patient survival were estimated using the Kaplan–Meier method and compared using the log-rank test. Serum amylase level was expressed as the geometric mean, and 95% CI and was compared between the non-ES and ES groups using a 2-way analysis of variance. All statistical analyses were performed using SPSS (version 19.0; IBM, Tokyo, Japan) and SAS (version 9.3; SAS, Tokyo, Japan) software. For all analyses, a P value less than .05 was considered statistically significant.
Results Patient Demographics and Self-Expandable Metallic Stent Placement–Related Data Between August 2010 and May 2012, there were 519 patients with biliary stricture caused by pancreatic cancer who were assessed for eligibility, and 319 patients were excluded. The 200 included patients were allocated randomly into equal-sized non-ES and ES groups. Biliary cannulation failure precluded SEMS placement in 2 and 4 patients in the non-ES and ES groups, respectively (Figure 1). Patient characteristics and SEMS
Figure 1. Flow diagram of patient enrollment, randomization, and follow-up evaluation.
June 2015
Sphincterotomy Before Stent Placement 1155
Table 2. Demographic and SEMS Placement-Related Data
Patient characteristic Age, mean SD, y Sex, male/female, n Initial drainage, n Endoscopic drainage (EBS/ENBD), n Percutaneous drainage (PTBD/PTGBD), n History of acute pancreatitis, n Gallbladder present, n Serum amylase level before the procedure, geometric mean (95% CI), IU/L MPD obstruction, n Maximum diameter of the proximal MPD, mean SD, mm SEMS placement–related data ERCP experience of endoscopists, median (range), y Type of duodenoscope (JF260V/TJF260V/TJF240/JF240/JF200; Olympus Medical Systems Corp, Tokyo, Japan), n Use of medical prophylaxis against post–SEMS placement pancreatitis (none/ulinastatin/gabexate mesilate/nafamostat mesilate/nonsteroidal anti-inflammatory drugs), n Guidewire inserted into the pancreatic duct, n Contrast medium injected into the pancreatic duct, n Extent of opacified pancreatic duct (main/branch/acinar), n Length of the biliary stricture, mean SD, mm Use of the rendezvous method, n SEMS length, 40/60/80/failure, mm, n Procedure time, mean SD, s
Non-ES (n ¼ 100)
ES (n ¼ 100)
P value
70.54 10.4 49/51 68 50/16 2/0 3 93 55.27 (48.19–63.38) 83 6.55 3.41
72.2 10.6 53/47 70 53/13 3/1 7 91 61.71 (53.21–71.56) 80 6.91 4.82
.27 .67 .88
9 (3–31) 52/43/2/0/3
9 (1–31) 51/45/1/1/2
.90 .92
28/44/17/11/0
31/45/17/7/0
.81
18 11 8/3/0 26.14 12.13 1 0/90/8/2 380.1 208.5
16 9 5/4/0 27.00 14.03 2 1/82/13/4 553.7 324.8
.85 .81
.21 .80 .28 .58 .54
.64 .62 .36 <.001
EBS, endoscopic biliary stenting; ENBD, endoscopic nasobiliary drainage; MPD, main pancreatic duct; PTBD, percutaneous transhepatic biliary drainage; PTGBD, percutaneous transhepatic gallbladder drainage.
placement–related data are summarized in Table 2. Patient characteristics were not significantly different between the non-ES and ES groups except for procedure time. Predictably, procedure time was significantly shorter in the non-ES group than in the ES group (387.9 203.3 vs 576.7 310.3 s; P < .001).
Rates of Early Adverse Events Specifically Related to the Presence or Absence of Endoscopic Sphincterotomy In the non-ES group, early AEs specifically related to the presence or absence of ES were observed in Table 3. Rates of Early AEs Specifically Related to the Presence or Absence of ES
Early AEs, n (%) Pancreatitis, n Mild/moderate/ severe, n Moderate bleeding, n Mild perforation, n
Non-ES (n ¼ 98)
ES (n ¼ 96)
9 (9.2) 8 3/3/2
10 (10.4) 9 6/3/0
0
1
1
0
Difference between non-ES and ES (95% CI) -1.2% (-9.6 to 7.1)
NOTE. Boldface indicates primary outcome measure of the study.
9 patients including post–SEMS placement pancreatitis in 8 patients and mild perforation during guidewire manipulation in 1 patient (Table 3). In the ES group, early AEs specifically related to the presence or absence of ES occurred in 10 patients, including post–SEMS placement pancreatitis in 9 patients and moderate bleeding in 1 patient (Table 3). Among the patients in whom SEMS placement was successful (on-treatment population), the percentage point difference in the rates of early AEs specifically related to the presence or absence of ES between the non-ES and ES groups was -1.2% (95% CI, 9.6% to 7.1%), indicating that non-ES was noninferior to ES. Furthermore, the rates for early and late SEMS placement–related AEs were not significantly different between the 2 groups in the ITT population (Supplementary Tables 1–3).
Change in Serum Amylase Level After Self-Expandable Metallic Stent Placement In the non-ES group, the serum amylase level increased significantly after SEMS placement from 54.91 IU/L (95% CI, 47.80–63.09 IU/L) to 116.97 IU/L (95% CI, 92.84–147.38 IU/L; P < .001). The serum amylase level also increased significantly after SEMS placement in the ES group from 63.14 IU/L (95% CI, 54.21–73.5 IU/L) to 125.76 IU/L (95% CI, 93.53–161.56 IU/L; P < .001). However, 2-way analysis of variance showed no significant difference in the change in serum amylase between
1156 Hayashi et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
Discussion
Figure 2. Change in serum amylase level after SEMS placement in the non-ES and ES groups. The large black circles represent the geometric mean of the serum amylase level, and the error bars indicate the associated 95% confidence interval.
the 2 groups in the ITT population (interaction P ¼ .38) (Figure 2).
Time to Self-Expandable Metallic Stent Dysfunction and Patient Survival Time The last patient was enrolled in May 2012, and the follow-up period ended in November 2012. The median follow-up duration was 192 days (range, 14–810 d) and 219.5 days (range, 14–757 d) in the non-ES and ES groups, respectively. The median time to SEMS dysfunction was not significantly different between the non-ES and ES groups in the ITT population (>594 vs 541 d; P ¼ .88). The median patient survival time also was not significantly different between the non-ES and ES groups in the ITT population (202 vs 255 d, respectively; P ¼ .20) (Supplementary Table 4, Figure 3).
We performed a noninferiority study to determine whether omitting ES before SEMS placement for distal biliary stricture caused by unresectable pancreatic cancer would increase the frequency of AEs specifically related to the presence or absence of ES (pancreatitis, bleeding, and perforation) and found that ES omission did not affect the rates of AEs specifically related to the presence or absence of ES. Thus, the present study overturns the conventional wisdom that ES prevents pancreatitis after SEMS placement. Generally, mechanical injury from manipulation of the papilla of Vater or the pancreatic duct is considered the major cause of post-ERCP pancreatitis. In addition to careful manipulation and adequate skill level of the endoscopist, nonsteroidal anti-inflammatory drugs and pancreatic stenting also can prevent pancreatitis. Pancreatic juice congestion caused by pancreatic orifice occlusion is considered another etiologic cause of postSEMS placement pancreatitis. ES theoretically could prevent such congestion by decreasing the pressure that the SEMS exerts toward the pancreatic orifice. However, the increase in serum amylase level, a common measure of pancreatic juice congestion, was similar between the non-ES and ES groups, indicating that ES does not protect the pancreatic duct orifice after SEMS placement. There are 2 possible explanations for the lack of protective effect: the expansive force of the SEMS may be so strong that ES cannot protect the pancreatic orifice, and/ or the axial (straightening) force of the SEMS may directly affect the pancreatic orifice rather than being limited to the incised duodenal mucosa as previously reported.19,20 The rates of perforation and bleeding in the ES group were only 0% and 1%, respectively, and were comparable with previously reported rates for
Figure 3. (A) Kaplan– Meier graph showing the time to SEMS dysfunction. (B) Kaplan–Meier graph showing patient survival.
June 2015
covered SEMS (0%–0.7% and 0%–2%, respectively).4,5,7–9,11–15 These rates are relatively low compared with those for ES for extraction of bile duct stones, and this may be attributable to the wrapping and tamponade effect of covered SEMS because the covering membrane may seal tiny perforations and other minor vessel injuries that occur during ES. A significant limitation of the present study was the unmasked assessment. It was impossible to mask the physicians (assessors) who evaluated the AEs to the treatment allocation because of the shortage of staff, especially in some of the small, rural, secondary referral hospitals. In general, failure to mask studies investigating the superiority of a newly developed procedure is thought to impose a high risk of substantial assessor bias. However, the purpose of this study was to show the noninferiority of non-ES relative to ES before SEMS placement. ES was a frequently performed procedure at the time the study was conducted; therefore, we would have expected any assessor bias to be slanted toward its presumed benefits. Therefore, we doubt that the unmasked assessment substantially affected our results. Another limitation was the undefined administration of prophylactics for pancreatitis. We cannot completely rule out the potential influence of prophylactic use on the pancreatitis rate. However, we believe that non-ES is noninferior to ES despite the use of prophylactics because the preventive effect of the prophylactics used in the present study has not been proven in any clinical study and the proportion of patients receiving the different types of prophylactics was not significantly different between the non-ES and ES groups. Our study also was not multinational. In prospective multicenter clinical studies in Japan, the rate of post-ERCP pancreatitis in high-risk and unselected subjects has been reported to be 11.3% and 3.5%, respectively.21,22 These rates are similar to those reported in other ethnic groups.23,24 In addition, one of the major causes of post–metallic stent placement pancreatitis is the occlusion of the pancreatic orifice. The occurrence of pancreatic orifice occlusion is not influenced by the endoscopist’s skill and may not be affected by patient race or ethnicity. We believe our study findings are applicable to all patients regardless of ethnicity, and thus, will impact clinical practice worldwide. In conclusion, the present study showed that ES does not affect the incidence of AEs, SEMS patency rate, or patient survival. Therefore, we conclude that ES is not necessary before SEMS placement for distal biliary stricture caused by unresectable pancreatic cancer.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2015.01.008.
Sphincterotomy Before Stent Placement 1157
References 1. Porta M, Fabregat X, Malats N, et al. Exocrine pancreatic cancer: symptoms at presentation and their relation to tumour site and stage. Clin Transl Oncol 2005;7:189–197. 2. Davids PH, Groen AK, Rauws EA, et al. Randomised trial of selfexpanding metal stents versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:1488–1492. 3. Knyrim K, Wagner HJ, Pausch J, et al. A prospective, randomized, controlled trial of metal stents for malignant obstruction of the common bile duct. Endoscopy 1993;25:207–212. 4. Soderlund C, Linder S. Covered metal versus plastic stents for malignant common bile duct stenosis: a prospective, randomized, controlled trial. Gastrointest Endosc 2006;63:986–995. 5. Isayama H, Yasuda I, Ryozawa S, et al. Results of a Japanese multicenter, randomized trial of endoscopic stenting for nonresectable pancreatic head cancer (JM-test): Covered Wallstent versus DoubleLayer stent. Dig Endosc 2011;23:310–315. 6. Artifon ELA, Sakai P, Ishioka S, et al. Endoscopic sphincterotomy before deployment of covered metal stent is associated with greater complication rate: a prospective randomized control trial. J Clin Gastroenterol 2008;42:815–819. 7. Isayama H, Kawabe T, Nakai Y, et al. Management of distal malignant biliary obstruction with the ComVi stent, a new covered metallic stent. Surg Endosc 2010;24:131–137. 8. Kullman E, Frozanpor F, Söderlund C, et al. Covered versus uncovered self-expandable nitinol stents in the palliative treatment of malignant distal biliary obstruction: results from a randomized, multicenter study. Gastrointest Endosc 2010;72:915–923. 9. Tsuchiya T, Itoi T, Gotoda T, et al. A multicenter prospective study of the short-term outcome of a newly developed partially covered self-expandable metallic biliary stent (WallFlex(®)). Dig Dis Sci 2011;56:1889–1895. 10. Aadam AA, Evans DB, Khan A, et al. Efficacy and safety of selfexpandable metal stents for biliary decompression in patients receiving neoadjuvant therapy for pancreatic cancer: a prospective study. Gastrointest Endosc 2012;76:67–75. 11. Isayama H, Mukai T, Itoi T, et al. Comparison of partially covered nitinol stents with partially covered stainless stents as a historical control in a multicenter study of distal malignant biliary obstruction: the WATCH study. Gastrointest Endosc 2012; 76:84–92. 12. Isayama H, Komatsu Y, Tsujino T, et al. A prospective randomized study of “covered” versus “uncovered” diamond stents for the management of distal malignant biliary obstruction. Gut 2004;53:729–734. 13. Costamagna G, Tringali A, Reddy DN, et al. A new partially covered nitinol stent for palliative treatment of malignant bile duct obstruction: a multicenter single-arm prospective study. Endoscopy 2011;43:317–324. 14. Telford JJ, Carr-Locke DL, Baron TH, et al. A randomized trial comparing uncovered and partially covered self-expandable metal stents in the palliation of distal malignant biliary obstruction. Gastrointest Endosc 2010;72:907–914. 15. Ung KA, Stotzer PO, Nilsson A, et al. Covered and uncovered self-expandable metallic Hanarostents are equally efficacious in the drainage of extrahepatic malignant strictures. Results of a double-blind randomized study. Scand J Gastroenterol 2013; 48:459–465. 16. Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37:383–393.
1158 Hayashi et al 17. Hirota M, Takada T, Kawarada Y, et al. Diagnostic criteria and severity assessment of acute cholecystitis: Tokyo guidelines. J Hepatobiliary Pancreat Surg 2007;14:78–82. 18. Wada K, Takada T, Kawarada Y, et al. Diagnostic criteria and severity assessment of acute cholangitis: Tokyo guidelines. J Hepatobiliary Pancreat Surg 2007;14:52–58. 19. Kawakubo K, Isayama H, Nakai Y, et al. Risk factors for pancreatitis following transpapillary self-expandable metal stent placement. Surg Endosc 2012;26:771–776. 20. Tarnasky PR, Cunningham JT, Hawes RH, et al. Transpapillary stenting of proximal biliary strictures: does biliary sphincterotomy reduce the risk of postprocedure pancreatitis? Gastrointest Endosc 1997;45:46–51. 21. Sofuni A, Maguchi H, Itoi T, et al. Prophylaxis of postendoscopic retrograde cholangiopancreatography pancreatitis by an endoscopic pancreatic spontaneous dislodgement stent. Clin Gastroenterol Hepatol 2007;5:1339–1346. 22. Kawakami H, Maguchi H, Mukai T, et al. A multicenter, prospective, randomized study of selective bile duct cannulation performed by multiple endoscopists: the BIDMEN study. Gastrointest Endosc 2012;75:362–372. 23. ASGE Standards of Practice CommitteeAnderson Michelle A, Fisher Laurel, et al. Complication of ERCP. Gastrointest Endosc 2012;75:467–473. 24. Dumonceau JM, Andriulli A, Elmunzer BJ, et al. Prophylaxis of post-ERCP pancreatitis: European Society of Gastrointestinal Endoscopy (ESGE) guideline - updated June 2014. Endoscopy 2014;46:799–815.
Reprint requests Address requests for reprints to: Tsuyoshi Hayashi, MD, Department of Medical Oncology and Hematology, Sapporo Medical University School of Medicine, South-1, West-16, Chuo-ku, Sapporo 060-8543, Hokkaido, Japan. e-mail: [email protected]; fax: (81) 11-612-7987.
Clinical Gastroenterology and Hepatology Vol. 13, No. 6 Acknowledgments The authors gratefully acknowledge the assistance of Michihiro Ono, MD, Makoto Yoshida, MD, Fumito Tamura, MD, Shingo Tanaka, MD, Hiroto Horiguchi, MD, and Junji Kato, MD (Department of Medical Oncology and Hematology, Sapporo Medical University); Masaki Kuwatani, MD, Kazumichi Kawakubo, MD, Kazunori Eto, MD, Shin Haba, MD, Taiki Kudo, MD, Yoko Abe, MD, Shuhei Kawahata, MD, Masahiro Asaka, MD, and Naoya Sakamoto, MD (Department of Gastroenterology and Hepatology, Hokkaido University Graduate School of Medicine); Hiroyuki Maguchi, MD, Kuniyuki Takahashi, MD, Akio Katanuma, MD, Kei Yane, MD, Shin Kato, MD, Maki Kaneko, MD, Shunpei Hashigo, MD, Ryo Harada, MD, and Ryusuke Kato, MD (Center for Gastroenterology, Teine-Keijinkai Hospital); Hiroaki Yamato, MD (Department of Gastroenterology, Hakodate Municipal Hospital); Hiroyuki Isshiki, MD, Kentaro Kawakami, MD, and Takafumi Naito, MD (Department of Gastroenterology, Muroran City General Hospital); Toru Kawamoto, MD, Junpei Sasajima, MD, and Yoshiaki Sugiyama, MD (Division of Gastroenterology and Hematology/ Oncology, Department of Medicine, Asahikawa Medical University); Naoki Uemura, MD (Department of Gastroenterology, Steel Memorial Muroran Hospital); Hiroyuki Miyakawa, MD (Department of Bilio-Pancreatology, Sapporo Kosei General Hospital); Takahiro Osuga, MD (Department of Gastroenterology, National Hospital Organisation Hokkaido Cancer Center); Ryusuke Matsumoto, MD (Third Department of Internal Medicine, Obihiro-Kosei General Hospital); Atsushi Chiba, MD (Department of Internal Medicine, Asahikawa City Hospital); Masahiro Shitani, MD (Department of Gastroenterology, Rheumatology, and Clinical Immunology, Sapporo Medical University); Keisuke Shinada, MD, and Akihiko Yumino, MD (Department of Gastroenterology, Keiwakai Ebetsu Hospital); Keisuke Yamakita, MD, and Masako Imazawa, MD (Division of Metabolism and Biosystemic Science, Department of Internal Medicine, Asahikawa Medical University); Manabu Onodera, MD (Department of Gastroenterology, Abashiri-Kosei General Hospital); Sho Takahashi, MD (Department of Gastroenterology, Otaru Kyokai Hospital); Tetsuya Miseki, MD (Department of Internal Medicine, Obihiro Daiichi Hospital); Minoru Uebayashi, MD (Department of Gastroenterology, Kitami Red Cross Hospital); Hideyuki Ihara, MD (The Center for Digestive Diseases, Tonan Hospital); Takuro Machida, MD, and Shuji Ikari, MD (Department of Internal Medicine, Hokkaido Gastroenterology Hospital); Nobuaki Akakura, MD (Department of Gastroenterology, Sapporo Medical Center Nippon Telegraph and Telephone East Corporation); Takehiro Kukitsu, MD (Department of Gastroenterology, Rumoi Municipal Hospital); and Kunihiro Takanashi, MD (Department of Gastroenterology, Otaru Ekisaikai Hospital). Conflicts of interest The authors disclose no conflicts. Funding Tsuyoshi Hayashi is supported in part by a grant from The Japanese Foundation for Research and Promotion of Endoscopy.
June 2015
Sphincterotomy Before Stent Placement 1158.e1
Supplementary Table 1. Rates of Early AEs Related to SEMS Placement
Early AEs, n (%) Pancreatitis, n Mild/moderate/severe, n Moderate bleeding, n Mild perforation, n Moderate liver abscess, n Epigastric pain, n Vomiting, n
Non-ES (n ¼ 100)
ES (n ¼ 100)
15 (15) 8 3/3/2 0 1 3 3 0
15 (15) 9 6/3/0 1 0 1 2 2
P value 1.0
Supplementary Table 2. Rates of Late AEs Related to SEMS Placement
Late AEs, n Severe pancreatitis, n Bleeding, n Mild/moderate/severe, n Moderate liver abscess, n Duodenal ulcer, n
Non-ES (n ¼ 100)
ES (n ¼ 100)
5 0 1 0/1/0 1 1
6 1 1 0/0/1 0 0
Supplementary Table 3. Rates of Early and Late-Onset Cholecystitis
P value 1.0
Early onset, n (%) Mild/moderate/severe, n Late onset, n (%) Mild/moderate/severe, n
Non-ES (n ¼ 93)
ES (n ¼ 91)
4 (4.3) 0/3/1 4 (4.3) 0/4/0
1 (1.1) 0/1/0 1 (1.1) 0/1/0
P value .37 .37
1158.e2 Hayashi et al
Clinical Gastroenterology and Hepatology Vol. 13, No. 6
Supplementary Table 4. Time to SEMS Dysfunction and Patient Survival Time
Time to SEMS dysfunction, median (95% CI), da Frequency of cholangitis, n (%) Cause of cholangitis, n (%) Food impaction Sludge formation Intestinal obstruction Proximal migration Distal migration Tumor ingrowth Tumor overgrowth Bleeding Unknown Patient survival, median (95% CI), db
10 5 2 2 1 1 3 0 1 202
Non-ES
ES
P value
>594 25
541 (353.4–728.7) 25
.88
(40) (20) (8) (8) (4) (4) (12) (4) (170.5–233.5)
10 4 3 3 1 2 0 1 1 255
(40) (16) (12) (12) (4) (8) (4) (4) (187.9–322.1)
.20
NOTE. In the non-ES group, the SEMS was removed in 7 patients. Fifty-two and 14 patients were censored at the time of patient death or at the end of the followup period, respectively. In the ES group, the SEMS was removed in 2 patients. Forty-nine and 20 patients were censored at the time of patient death or the end of the follow-up period, respectively. SEMS dysfunction developed in 25 patients in each group. The median time to SEMS dysfunction could not be calculated because it was longer than 594 days in the non-ES group and was 541 days in the ES group as estimated by the Kaplan–Meier method; this difference was not significant (P ¼ .88). The main causes of cholangitis were food impaction and sludge formation. The frequencies of all causes of SEMS dysfunction were similar between the 2 groups. Death from pancreatic cancer progression occurred in 78 patients in the non-ES and in 67 in the ES group. The median patient survival time was 202 days in the non-ES group and 255 days in the ES group, a nonsignificant difference (P ¼ .20). a The median (95% CI) time to SEMS dysfunction was estimated by the Kaplan–Meier method. b The median (95% CI) patient survival time was estimated by the Kaplan–Meier method.