Introduction

Malignant strictures are one of the causes of the biliary tree obstruction, apart from inflammatory, autoimmune, or iatrogenic ones. Most commonly, they are caused by adenocarcinoma of the head of the pancreas, cholangiocarcinoma, gallbladder cancer with extension into the common bile duct, tumors of the ampulla of Vater, and metastases (usually of gastrointestinal tract cancer or breast cancer) to the hepatoduodenal ligament.1,2,3

Medical treatment of a patient with malignant obstruction of the bile ducts serves only as a temporary measure, and long-term relief of the biliary obstruction is necessary to prevent an adverse outcome.4

One of the methods of jaundice treatment in the course of the disease, apart from percutaneous or surgical biliary drainage, is endoscopic biliary stenting with self-expandable metal stents (SEMSs) or plastic stents (PSs).5

Endoscopic biliary stenting is the most cost-effective method of treatment, and when biliary drainage is indicated, endoscopic stenting is considered the first-line technique of choice.6 According to Artifon et al,7 mean (SD) total costs of endoscopic SEMS placement were approximately half of those of surgery (4271 [2411] USD vs 8321 [1821] USD).7 Similarly, in an analysis of a Unites States database, Inamdar et al8 revealed a lower rate of adverse events and lower total costs of endoscopic retrograde cholangiopancreatography (ERCP) than of percutaneous transhepatic biliary drainage.8

The commonly used PSs are double-pig-tail (DPT) or straight in shape, and SEMSs are available as fully covered, partially covered, or uncovered. Metal stents are most often made of nitinol, which is an alloy of nickel and titanium.9,10 Patency of a stent and incidence of adverse events associated with the procedure affect the quality of patient life. Stent dysfunctions include stent migration, stent occlusion due to a tumor over- and / or ingrowth, and encrustation (sludge). The idea behind implementation of coating for SEMSs was to prevent tumor ingrowth and to allow endoscopic removal of the stents. According to some authors, disadvantages of covered stents include more frequent migration, occlusion by biliary sludge, and an increased risk of acute pancreatitis.11 As the uncovered stents are nonremovable, they are used as palliative treatment in the cases of histopathologic confirmation of cancer or indisputable malignant nature of the stricture. Their advantage is that they can be used in the strictures within the area of the hepatic hilum. Some studies have investigated differences between covered and uncovered stents in terms of their functioning time, occurrence of complications, and patient survival time, but the results differ widely.12,13 Moreover, these meta-analyses focused on complications, and did not pay attention to differences in patient survival and incidence of adverse events and dysfunction events.

Endoscopic biliary stenting during ERCP is associated with a high complication rate, and therefore the procedure requires careful qualification. The most common complications include acute pancreatitis (3.5%–9.7%), cholangitis and cholecystitis (1.4%), bleeding (1.3%), and perforation (0.6%).14

Biliary stenting can be used as a method of palliative treatment or as a bridge before surgical treatment, although it is not routinely recommended in the case of a quick access to surgery.15 Palliative treatment is an important therapy, as in about 70% of patients the malignant distal biliary strictures are nonresectable at the moment of diagnosis.16 Preoperative biliary drainage in patients with jaundice should only be considered in specific situations, which include coagulopathy, acute kidney injury, persistent itching, bacterial cholangitis, systemic inflammation, and long waiting time for the operative procedure as well as scheduled neoadjuvant chemotherapy / radiochemotherapy in icteric patients.17,18 In patients with pancreatic cancer noneligible for surgery, biliary stenting allows for prolonging the survival time and is a considerably more effective treatment than percutaneous biliary drainage.19 In the guidelines from 2018,20 the European Society of Gastrointestinal Endoscopy (ESGE) recommended the use of SEMSs in patients with malignant biliary strictures, as they showed advantages over PSs in terms of survival time, quality of life, number of reinterventions, and complications.21-24 Similar recommendations were listed in the American guidelines published in 2023.25 In clinical practice, PSs are exclusively used in patients without histopathologic evidence of cancer, with limited life expectancy, and in the case of strictures localized in the hepatic hilum. On the other hand, SEMSs are frequently underutilized due to higher initial costs of the procedure.

The aim of the study was to compare the effectiveness of SEMSs and PSs in the endoscopic drainage of malignant obstruction of bile ducts and its sequels for future optimization of this treatment method.

Patients and methods

The study included 618 consecutive patients with malignant biliary stricture subjected to 1271 procedures of biliary stenting in the years 2012 to 2017. A retrospective assessment of the implanted stent performance was conducted with a follow-up period of 3 years. Indications for stenting were malignant strictures confirmed by pathomorphologic examination of specimens collected during ERCP, endoscopic ultrasound, or percutaneous biopsy. The study comprised also 20 patients whose imaging and endoscopic examinations as well as clinical course of the disease clearly indicated its malignant nature. Depending on the indication for stenting, the patients were divided into 5 groups: ampullary cancer, pancreatic cancer, gallbladder cancer, cholangiocarcinoma, and others (hepatocellular carcinoma, less common types of tumors [eg, lymphoma], and metastases to the hepatoduodenal ligament and the liver). Regarding the level of obstruction, ampullary and pancreatic cancers were considered “low,” whereas the remaining tumors (gallbladder cancer, cholangiocarcinoma, and others) were considered “high.” Stent dysfunction was diagnosed according to the criteria proposed by Schmidt et al,26 which include biliary dilatation on ultrasound (as compared with the bile duct diameter measured after stenting), bilirubin concentration equal to or above 2 mg/dl or more than 1 mg/dl difference, as compared with that recorded after successful stenting, at least a double increase in alkaline phosphatase (ALP) or γ-glutamyl transpeptidase (GGT) activity above the upper limit or an increase by at least 30 IU/l and symptoms of cholangitis (fever or leukocyte count >10 000/mm3, or C-reactive protein [CRP] level >20 mg/l).

Routinely determined laboratory parameters (bilirubin, aspartate aminotransferase, alanine aminotransferase [ALT], GGT, ALP, CRP) were analyzed to compare the efficiency of biliary drainage. Laboratory tests performed just before and after the procedure (just before discharge, but not later than 10 days after stenting) were included in the analysis.

ERCP procedures were performed by 5 experienced endoscopists (>2000 procedures per endoscopist), using Olympus TJF-145, TJF-160VR, or Olympus TJF-Q180V endoscopes (Olympus, Hamburg, Germany). All patients signed their informed consent before the procedure.

The procedures were performed under deep intravenous sedation. For prophylaxis of post-ERCP pancreatitis, the patients were routinely given anal suppository of 100 mg of diclofenac immediately before or immediately after ERCP and adequate intravenous hydration. Antibiotic prophylaxis was used according to the ESGE recommendations.14

A decision on the type of prosthesis to be inserted (PS vs SEMS, uncovered vs covered SEMS, DPT vs straight PS, manufacturer, model, dimensions, etc.) was made by the endoscopist based on clinical data and the anatomic conditions found during ERCP in accordance with the ESGE guidelines. As a rule, nonremovable (uncovered or partially covered) SEMSs were inserted only in the patients with histologically confirmed cancer. For lesions located at the confluence of hepatic ducts, only PSs or uncovered SEMSs were used. To obtain the best drainage, we attempted to insert stents of the largest possible diameter, that is, 10 F (outside diameter; 1 French [F] equals 0.33 mm) for PS and 10 mm for SEMS. The most frequently used SEMS model (48%) was a partially covered Wallstent (Boston Scientific, Galway, Ireland), followed by an uncovered Wallstent (Boston Scientific) (43%), and fully covered WallFlex (Boston Scientific) (5%). In 6 patients (2%) with type II to IV Bismuth–Corlette cholangiocarcinoma and attempted bilateral Y-configuration, wide-mesh stents Hanarostent Hilar (M.I.Tech, Pyeongtaek, Republic of Korea) and Niti-S LCD stents (TaeWoong Medical, Gimpo, Republic of Korea) were used. All DPT PSs were of PE 204 series (Endo-Flex, Voerde, Germany), whereas all straight PSs were of PBD series (Olympus Europa, Hamburg, Germany).

For simplified cost analysis, it was assumed that the use of resources during ERCP was identical for the insertion of PSs and SEMSs, except for the stents and additional equipment necessary for PS insertion (pusher and guiding catheter or stent insertion kit). The current prices of SEMSs and PSs (including insertion kits) were obtained from 6 major providers of endoscopic equipment, while the data on current reimbursement values were obtained from the Polish National Health Fund (Narodowy Fundusz Zdrowia [NFZ]). It was also assumed that any other costs (drugs, laboratory tests, etc.) were equal in both groups.

According to the Ethics Committee, this study did not need ethical approval.

Statistical analysis

Statistical calculations were made with Statistica 12.0 package (StatSoft, Kraków, Poland). Data for continuous variables were expressed as median (interquartile range [IQR]) or mean (SD). Qualitative variables were presented as numbers (percentage). The Mann–Whitney test was used for the comparison of continuous variables, whereas the χ2 test was used to compare qualitative variables. Differences were deemed significant at P <⁠0.05.

Results

General data

The analysis included 1271 procedures of biliary stenting conducted in 618 patients with malignant strictures. PSs were used in 929 procedures (73%), of which 637 stents (69% of PSs) were of the straight type; 88% of them were 10 F, 9% 8.5 F, and 3% 7 F stents. Out of 292 (31% of PSs) DPT stents, 88% were 7 F, 4% 8.5 F, and 8% 10 F. SEMSs were used in 342 procedures (27%); 156 of them (46%) were uncovered, while 185 (54%) were partly (167; 49%) or fully (19; 5%) covered.

Patient characteristics, indications for stenting, and length of hospitalization are presented in Table 1. Median patient age was 67.5 years (IQR, 59.4–75.1), and 55% were women. The age and sex distribution in the PS and SEMS groups was similar.

Table 1. Clinical and procedural characteristics of patients

Variable

Total

Type of stent

P value

PS

SEMS

Patients, n

618

456

162

–

Stenting procedures, n

1271

929

342

–

Age, y

Mean (SD)

67 (11.1)

67.1 (11.1)

66.7 (11)

0.95

Median (IQR)

67.5 (59.4–75.1)

67.5 (59–75.2)

67.7 (60.1–74.3)

Sex, n (%)

Women

341 (55.2)

249 (54.6)

92 (56.8)

0.63

Men

277 (44.8)

207 (45.4)

70 (43.2)

Indication for stenting, n (%)

Ampullary cancer

140 (11)

115 (12.4)

25 (7.3)

<⁠0.001

Pancreatic cancer

468 (36.8)

301 (32.4)

167 (48.8)

Gallbladder cancer

138 (10.9)

99 (10.7)

39 (11.4)

Cholangiocarcinoma

432 (34)

354 (38.1)

78 (22.8)

Other

93 (7.3)

60 (6.5)

33 (9.6)

Level of obstruction, n (%)

Low (ampullary and pancreatic cancer)

608 (47.8)

416 (44.8)

192 (56.1)

0.001

High (other indications)

663 (52.2)

513 (55.2)

150 (43.9)

Hospital stay, d

Mean (SD)

11.2 (7.4)

11.8 (7.9)

9.5 (5.6)

<⁠0.001

Median (IQR)

9 (6–14)

10 (6–15)

8 (6–11)

Abbreviations: IQR, interquartile range; PS, plastic stent; SEMS, self-expandable metal stent

There were more pancreatic cancers in the SEMS group (49% SEMS vs 32% PS), and more cholangiocarcinoma cases in the PS group (38% PS vs 23% SEMS). Consequently, there were more cases of high-level obstruction in the PS group (55% high vs 45% low), while an opposite proportion was observed in the SEMS group (44% high vs 56% low).

Patency

In Table 2 we present detailed data on stent patency in the PS and SEMS groups as well as for different types of PSs (straight vs DPT) and SEMSs (uncovered vs covered) for different indications and obstruction levels.

Table 2. Stent patency vs indication for stenting and type of stent

Compared variable

Stent patency, d

PS vs SEMS

Type of PS

Type of SEMS

Stent type

PS

SEMS

P value

DPT

Straight

P value

uSEMS

cSEMS

P value

All procedures

46 (18–97)

118 (56–232)

<⁠0.001

34.5 (14–73)

52 (21–206)

0.001

107 (57–219)

131 (55–242)

0.71

Level of obstruction

Low

53.5 (20–107)

129.5 (59–247)

<⁠0.001

28 (9–54)

58 (24–111)

0.002

122.5 (65–239)

132 (56–256)

0.82

High

41 (17–91)

107 (54–218)

<⁠0.001

38 (16–77)

42.5 (18–96)

0.08

103 (55–214)

125 (48–237)

0.79

Indication for stenting

Ampullary cancer

43(10–102)

177 (105–277)

<⁠0.001

35 (13–104)

44 (10–102)

0.97

22.5 (5–40)

201 (136–347)

0.03

Pancreatic cancer

57 (24–111)

116 (56–233)

<⁠0.001

21 (8–49)

63 (28–119)

<⁠0.001

127.5 (67–244)

107.5 (54–223)

0.2

Gallbladder cancer

35 (16–82)

79 (30–162)

0.005

28.5 (12–48)

42 (16–92)

0.13

68 (24–110)

85 (33–226)

0.42

Cholangiocarcinoma

42 (18–91)

115.5 (62–254)

<⁠0.001

41 (17–86)

47 (18–95)

0.35

109 (62–223)

138 (66–269)

0.56

Other

41(16–113)

136 (48–219)

0.007

31.5 (16–56)

50 (17–136)

0.35

142.5 (43–221)

134 (48–166)

0.81

Data are expressed as median (interquartile range).

Abbreviations: cSEMS, covered SEMS; DPT, double pig-tail; uSEMS, uncovered SEMS; others, see Table 1

SEMSs were patent for longer than PSs (mean [SD], 118 [56–232] days vs 46 [18–97] days; P <⁠0.001). Superiority of SEMSs was observed for all indications and for both (low and high) obstruction levels.

In the PS group, straight stents exhibited a longer median patency time than DPT stents (52 [21–106] days vs 34.5 [14–73] days; P = 0.001). For individual indications straight stents were better than DPT stents in the patients with pancreatic cancer (63 vs 21 days; P <⁠0.001) and, consequently, in all cases with a low-level obstruction (58 vs 28 days; P = 0.001). For all other indications and high-level obstruction the differences were insignificant. In the SEMS group, we did not find any clinically important significant differences between uncovered and covered stents.

A decreased median stent patency time was observed in the patients with high vs low obstruction level (49 [20–109] days vs 66 [28–136] days; P = 0.003). In the subgroup analysis, this effect was present in the PS group (41 vs 53.5 days; P = 0.04), and it was abolished in the patients treated with SEMSs (107 vs 129.5 days; P = 0.26).

Laboratory results

Effectiveness of biliary drainage measured by laboratory tests was also better in the patients treated with SEMSs (Table 3). A relative decrease, as compared with baseline, of bilirubin concentration as well as ALP and ALT activity was significantly greater in the SEMS than in the PS group. Median (IQR) bilirubin drop was 37% (14%–58%) in the SEMS group and 32% (9%–53%) in the PS group (P = 0.01). Moreover, the patients in the SEMS group showed a CRP level decrease by 10.5%, whereas in the PS group the CRP level rose by 17% (P = 0.02).

Table 3. Laboratory parameters before and after stenting with plastic stents and self-expandable metal stents

Variable

All procedures

Type of stent

P value

PS

SEMS

Bilirubin, mg/dl

Before

14.1 (7.9–22.7)

15.8 (9.6–24.4)

11.5 (5.7–20.6)

0.01

After

8.6 (3.2–17.2)

10.8 (4.4–19.5)

6.8 (2.8–13.7)

Change (% of baseline)

–33.9 (–55.7 to –9.5)

–31.5 (–53.3 to –8.6)

–37.2 (–57.5 to –13.9)

ALP, U/l

Before

506 (338–784)

474 (331–714)

558 (361–902)

0.02

After

405 (291–597)

386 (293–565)

438 (285–680)

Change (% of baseline)

–20.1 (–31.6 to –6.9)

–18.8 (–29.8 to –6.3)

–23.8 (–34.2 to –7.7)

GGT, U/l

Before

695 (359–1125)

659 (311–1082)

734 (396–1267)

0.92

After

499 (225–824)

490 (208–767)

531 (233–911)

Change (% of baseline)

–28.3 (–42 to –11.1)

–27.1 (41.8 to –12.6)

–29.0 (–43.6 to –9.3)

ALT, U/l

Before

128 (69–238)

143 (73–252)

111 (64–210)

0.03

After

79 (45–138)

86 (53–150)

69 (40–120)

Change (% of baseline)

–34.8 (–55.1 to –14.5)

–32.8 (–51.1 to –12.7)

–38.2 (–57.5 to –20.8)

CRP, mg/l

Before

23.9 (11.8–74.6)

17.6 (9.2–45.9)

39.2 (16.3–99.8)

0.02

After

29.8 (14.3–71.5)

23.6 (12.1–60.9)

36.5 (16.5–83.6)

Change (% of baseline)

7.6 (–42 to 91.6)

17.0 (–26.1 to 102.2)

–10.5 (–55.2 to 67.9)

Data are expressed as median (interquartile range).

SI conversion factors: to convert ALP, ALT, and GGT to μkat/l, multiply by 0.0167; bilirubin to μmol/l, by 17.104.

Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; CRP, C-reactive protein; GGT, γ-glutamyl transpeptidase; others, see Table 1

Complications

The most frequent stent-related complications included distal partial migration (5.2%), acute pancreatitis (2.8%), and liver abscess formation (2%). The use of PSs was associated with a higher risk of complications during the observation period, as compared with SEMSs (19.3% vs 12.9%; P = 0.001). A detailed analysis of complications is given in Table 4.

Table 4. Stent-related complications (per procedure)

Complication (% of all complications)

All procedures, n (%)

Type of stent

P value

PS, n (%)

SEMS, n (%)

Distal migration (partial) (29.6)

66 (5.2)

59 (6.4)

7 (2)

0.001

Acute pancreatitis (16.1)

36 (2.8)

23 (2.5)

13 (3.8)

0.25

Liver abscess (11.2)

25 (2)

20 (2.2)

5 (1.5)

0.50

Bleeding (10.8)

24 (1.9)

16 (1.7)

8 (2.3)

0.49

Proximal migration (9.4)

21 (1.7)

15 (1.6)

6 (1.8)

0.81

Complete migration (7.6)

17 (1.3)

17 (1.8)

0

0.001

Perforation (3.6)

8 (0.6)

7 (0.8)

1 (0.3)

0.69

Cholecystitis (2.2)

5 (0.4)

3 (0.3)

2 (0.6)

0.62

Other (9.4)

21 (1.7)

19 (2)

2 (0.6)

0.08

Total

223 (17.5)

179 (19.3)

44 (12.9)

0.001

Abbreviations: see Table 1

Costs

An average price of a SEMS was 990 USD (range for 6 suppliers, 638–1348 USD), while an average price of a PS kit was 113 USD (range for 5 suppliers, 58–150 USD). Reimbursement from the NFZ for hospitalization lasting more than 2 days (>99% of the cases in our study) with SEMS insertion (International Classification of Diseases, Ninth Edition [ICD-9] procedure code, 51.872) was 3760 USD, while reimbursement for PS insertion (ICD-9 procedure code, 51.871) was 1951 USD (a difference of 1809 USD). Thus, an average profit for the hospital for the initial procedure was 932 USD when inserting SEMSs, ranging from 605 to 1228 USD for individual suppliers. Shortening of the hospital stay by 2.3 days further increased the profit for the hospital by approximately 443 USD (assuming the cost of 1 day of the hospital stay in a gastroenterology / general surgery ward at 155 and 230 USD, respectively). Thus, a total profit for the hospital for 1 stenting procedure with SEMS was approximately 1375 USD.

Assuming the identical survival rate in both groups, with patency time approximately 2.5 times longer in the SEMS than in the PS group, the costs for health care system are lower in the SEMS group (1 procedure × 3760 = 3760 USD vs 2.5 procedures × 1951 USD = 4877 USD; respectively), with savings for the system of 1117 USD per patient.

Discussion

The ESGE recommends SEMSs for treatment of malignant biliary strictures. In our study, SEMSs were associated with a longer patency time and shorter hospital stay than PSs.

The presence of coating did not have a significant influence on the patency time of SEMSs. Nevertheless, in all types of cancer except for pancreatic cancer, fully-covered stents exhibited longer patency, which is consistent with the results of Yamashita et al.27 In that meta-analysis of 2358 patients, the time to recurrent biliary obstruction was significantly longer for covered SEMSs (mean differences, 45.51 days; 95% CI, 11.79–79.24) than for uncovered ones. In the subgroup of patients treated with PSs, the straight stents had longer patency, which is most likely related to their larger diameter than that of the DPT stents (usually 10 F vs 7 F).

Tumors located in the hepatic hilum are characterized by worse prognosis and are associated with a risk of unsuccessful procedure, partly because they often require stenting of both hepatic ducts.28-31 In general, PSs implanted due to hilar area cancer have significantly shorter patency than the stents implanted due to distal strictures. Our study showed that an endoscopic reintervention after PS insertion can be expected after approximately 35 days for gallbladder cancer and 47 days for cholangiocarcinoma, whereas for tumors located in the distal part of the biliary tract this period is almost twice longer.32 Similar dependencies between the patency time and location were demonstrated in the case of SEMSs; however, we did not observe statistical significance in this group. The analysis of stent patency in individual types of cancer showed that patency time of both PSs and SEMSs was the shortest in the case of gallbladder cancer and cholangiocarcinoma. A faster decline in bilirubin concentration after SEMS insertion may be important in clinical practice, especially in patients qualified for palliative or neoadjuvant chemotherapy, in whom bilirubin concentration equal to or below 2 mg/dl is required to initiate treatment in most protocols.33

In our analysis of stent-related complications, we found more common distal and complete migrations of PSs than of SEMSs (19.3% vs 12.9%; P = 0.001). This finding could be expected due to a smaller diameter of PSs and their poor adhesion to the bile duct wall. We registered no cases of migration of uncovered SEMSs, which is most likely due to their stabilization by overgrowth with the biliary epithelium. The most frequent complication related to the use of covered stents was their partial distal migration, which is in accordance with published data.34,35 In our experience, the use of covered SEMSs was not associated with a higher rate of gallbladder empyema, acute pancreatitis, or liver abscesses found by other researchers.36 However, we did not verify if the patients underwent cholecystectomy before inclusion to the study, which could potentially impact the cholecystitis rate.

In their meta-analysis, Moss et al37 proved that due to a lower number of reinterventions the use of SEMSs is cost-efficient despite their higher price. Median incremental cost-effectiveness ratio of SEMS was 1820 USD per ERCP prevented. However, there are also meta-analyses that did not prove economic benefit of SEMSs.38 Walter et al39 studied the cost efficacy of SEMSs vs PSs in patients with inoperable extrahepatic malignant bile duct obstruction and found that the total cost after 1 year did not differ. It should be emphasized that advantages of SEMSs are abolished if used in patients with an advanced metastatic disease and prognosed survival time shorter than 3 months.39 In such cases, the use of PSs may be justified to reduce the costs of hospitalization without a significant impact on the clinical outcome.

Our study has several limitations related to its retrospective design. First, the stent type was chosen according to the endoscopists’ preferences, which can lead to a selection bias. We found a greater proportion of cholangiocarcinoma cases in the PS group, whereas pancreatic cancer was the dominant diagnosis in the SEMS group. The most likely explanation is that well-known difficulties in histologic confirmation of cholangiocarcinoma lead to more frequent PS insertion, as only uncovered, nonremovable SEMSs can be placed through the hepatic duct confluence. On the other hand, histologic confirmation of pancreatic cancer is much easier, and such patients can be treated with fully covered, removable SEMSs.

We have presented one of the largest series that evaluates the results of endoscopic stenting of malignant biliary strictures with 3-year follow-up. Our data suggest that SEMSs should be used in malignant strictures, and an exception from this rule are patients with disseminated cancer and very short life expectancy. The insertion of SEMSs results in better biliary drainage, longer patency time, and shorter hospitalization. The use of covered SEMSs was not associated with an increased risk of acute pancreatitis, cholecystitis, or liver abscesses, though the risk of distal migration must be considered. Finally, the cost analysis favors the insertion of SEMSs, when clinically appropriate.