Treatment of malignant gastroduodenal obstruction with a nitinol self-expanding metal stent: An international prospective multicentre registry

Page 1

D

Ts

GAMJa

b

c

d

e

f

g

h

i

j

k

l

a

ARAA

KDEGM

1

n

A

a

1d

Digestive and Liver Disease 44 (2012) 37– 43

Contents lists available at SciVerse ScienceDirect

Digestive and Liver Disease

j our nal ho me page: www.elsev ier .com/ locate /d ld

igestive Endoscopy

reatment of malignant gastroduodenal obstruction with a nitinolelf-expanding metal stent: An international prospective multicentre registry

uido Costamagnaa,∗, Andrea Tringali a, Julius Spicakb, Massimiliano Mutignania, John Shawc,ndre Royd, Erik Johnssone, Eduardo G.H. De Mouraf, Spencer Chengf, Thierry Ponchong,ax Bittingerh, Helmut Messmannh, Horst Neuhaus i, Brigitte Schumacher i, Rene Laugier j,

uha Saarniok, Francisco Igea Ariqueta l

Policlinico A. Gemelli, Rome, ItalyIKEM, Prague, Czech RepublicGroote Schuur Hospital, Cape Town, South AfricaCentre Hospitalier de Lúniversité de Montreal (CHUM), Montreal, CanadaSahlgren’s Hospital, Göteborg, SwedenHospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, BrazilEdouard Herriot, Lyon, FranceIII. Med. Klinik, Klinikum Augsburg, GermanyEvangelisches Krankenhaus Duesseldorf, Dusseldorf, GermanyHôpital de la Timone, Marseille, FranceOulu University Hospital, Oulu, FinlandHospital Rio Carrion, Palencia, Spain

r t i c l e i n f o

rticle history:eceived 15 April 2011ccepted 16 August 2011vailable online 19 September 2011

eywords:uodenal stentndoscopic stentingastric Outlet Obstruction Scorealignant gastric and duodenal stricture

a b s t r a c t

Background: Duodenal stenting has become a broadly accepted first line of treatment for patients withadvanced malignant gastroduodenal obstruction as these patients are difficult to treat and are poorsurgical candidates.Aims: To document duodenal stent performance for palliative management of malignant gastroduodenalobstruction.Methods: Multicentre, single arm, prospective registry documenting peroral endoscopic duodenal stent-ing procedures in 202 patients.Results: Technical success achieved in 98% (CI, 95%, 99%) of stent placements. Increase of Gastric OutletObstruction Score by at least 1 point compared to baseline was achieved in 91% (CI, 86%, 95%) of patientspersisting for a median of 184 days (CI, 109, 266). By day 5 (CI, 4, 6) after stent placement, 50% of patientsexperienced a score increase of at least 1 point.

Improvement from 14% of patients at baseline tolerating soft solids or low residue/normal diet to 84%at 15 days, 86% at 30 days, 81% at 90 days, 79% at 180 days, and 70% at 270 days. Complications includedstent ingrowth and/or overgrowth (12.4%), transient periprocedural symptoms (3%), bleeding (3%), stentmigration (1.5%), and perforation (0.5%).Conclusions: Safety and effectiveness of duodenal stenting for palliation of malignant gastroduodenalobstruction was confirmed in the largest international prospective series to date.

Gast

© 2011 Editrice

. Introduction

Duodenal stenting has become a first line of treatment for malig-ant gastroduodenal obstruction. Patients indicated for duodenal

∗ Corresponding author at: Digestive Endoscopy Unit, Catholic University, Largo. Gemelli 8, 00168 Rome, Italy. Tel.: +39 06 30156580; fax: +39 06 30156581.

E-mail addresses: [email protected] (G. Costamagna),[email protected] (A. Tringali).

590-8658/$36.00 © 2011 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevieroi:10.1016/j.dld.2011.08.012

roenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

stenting frequently have advanced disease and are poor surgicalcandidates. Palliation of gastric outlet obstruction symptoms andimprovement of oral intake are the primary goals in these cases.Recent studies have documented high levels of clinical successand acceptable complication profiles using an uncovered metalstent for treatment of gastroduodenal obstruction [1–4]. The use

of uncovered metal stents has been associated with less need forreintervention [5].

Compared to open and laparoscopic surgery, endoscopic stent-ing is cost effective, as it usually provides a lesser procedure cost,

Ltd. All rights reserved.

Page 2

3 e and L

sAa

StsPp

2

2

iS

wgari

lrht

2

Wpsaiawatwcugs

tdesswdetip

2

2

io

8 G. Costamagna et al. / Digestiv

horter recovery time, and reduced length of hospitalisation [6–11].dditionally, surgery in patients with advanced disease has beenssociated with increased morbidity and mortality [7–9].

This international multicentre registry, sponsored by Bostoncientific Corporation (Natick, MA, USA), is the largest prospec-ive duodenal stenting series to date and aims to document theafety and effectiveness of a new Nitinol self expanding metal stent.articipants in the registry include mainly tertiary referral centreserforming duodenal stenting procedures per standard of practice.

. Materials and methods

.1. Device

The duodenal stent used was a Nitinol, uncovered, self expand-ng metal stent (SEMS) (WallFlexTM Enteral Duodenal Stent, Bostoncientific Corporation, Natick, MA, USA).

The stent is pre-loaded into a through the scope/over theire (TTS/OTW) delivery system that accommodates a 0.035 in.

uidewire. The 10 French delivery system may be inserted through minimum 3.7 mm endoscopic working channel. The stent can bee-sheathed if deployed less than 75%, so adjustments can be madef the position is not correct.

The Wallflex is composed of Nitinol wire braided in a tubu-ar mesh configuration resulting in a flexible, self expanding, andadiopaque stent. Stent lengths are 6 cm, 9 cm, or 12 cm. All stentsave a body diameter of 22 mm and a flare diameter of 27 mm athe proximal end.

.2. Procedure

Stent placement was performed in accordance with theallFlexTM Enteral Duodenal Stent Directions for Use. A thera-

eutic gastroscope or a side viewing duodenoscope was used fortent placement. The gastroduodenal stricture was passed with

guidewire and a standard ERCP catheter. Contrast was injectedmmediately downstream from the stricture to assess morphologynd length of the stricture. The stricture was measured by a catheterith radiopaque markers. Patency of the proximal jejunum was

lso evaluated by contrast injection to exclude the presence of stric-ures distal to the intended stent location. When the guidewireas correctly positioned distal to the stricture stent deployment

atheter was advanced over the wire and the stent was releasednder fluoroscopic and endoscopic control. Where possible theuidewire was positioned with the tip at the angle of Treitz fortable positioning.

The length of the stent was chosen to be at least 4 cm longerhan the length of the stricture allowing for possible further tumourevelopment and post implant stent shortening due to continuedxpansion. Choosing a stent using this method assures that thetent covers areas of disease free lumen at each end and that thetent can traverse angulated anatomy. Pre-stenting balloon dilationas not routinely used due to the risk of perforation. Post stentingilation was not routinely performed since Nitinol stents tend toxpand fully after 48–72 h. Correct positioning and performance ofhe stent were assessed immediately after deployment by contrastnjection through the scope. Fig. 1 provides illustrations of stentlacement.

.3. Study design

.3.1. The registryThe objective of the registry was prospectively to document clin-

cal experience with duodenal stenting for the palliative treatmentf malignant gastroduodenal obstruction.

iver Disease 44 (2012) 37– 43

The study was a multicentre international internet based reg-istry. There were 12 participating centres in 10 countries: Brazil,Canada, Czech Republic, Finland, France, Germany, Italy, SouthAfrica, Spain, and Sweden. Stents were purchased directly by eachinstitution. Eleven were tertiary referral centres and one was acommunity based hospital.

Clinical data were encrypted for transmission on the internetusing the SSL Protocol (Secure Socket Layout). Access was pass-word controlled and patient confidentiality was ensured. A clinicalstudy agreement was in place between the study stent manufac-turer and each participating institution. The study protocol andinformed consent forms were approved by the ethics committeeof the coordinating centre (Policlinico Gemelli, Rome, Italy) on 15December 2005.

The aim of the registry was to enrol 200 patients. Sample sizewas not based on formal hypothesis testing. Rather the objectiveof the registry was to provide documentation of safety and effec-tiveness of duodenal stenting in a wide range of settings with anopportunity for participating centres to contribute approximately10–20 patients each.

2.3.2. PatientsPatients were enrolled between August 2006 and November

2008. Patients were indicated for palliative treatment of malig-nant gastroduodenal obstruction. Exclusion criteria were enteralischemia, multiple strictures, suspected or impending perforation,perforation, intra-abdominal abscess, and/or contraindication forendoscopic examination. Patient enrollment by institution is asfollows: 64 Rome, 31 Prague and Cape Town, 17 Montreal, 16 Gote-borg, 14 Sao Paulo, 9 Lyon, 6 Augsburg and Dusseldorf, 5 Marseille,4 Oulu, 3 Palencia.

2.3.3. Study visitsDuration of post stent placement follow-up was 9 months.

Follow-up visits were conducted via telephone or as in-hospital vis-its at 15 days, 1, 3, 6, and 9 months post stent placement. Additionalunscheduled visits were performed as necessary.

2.3.4. EndpointsThe primary endpoint, clinical success, was defined as effective-

ness of the duodenal stent in providing relief of obstruction asmeasured by oral intake. Oral intake was assessed using the GastricOutlet Obstruction Scoring System based on a four point system:(0) no oral intake, (1) liquid only, (2) soft solids, or (3) low residueor normal diet [12]. Primary endpoint analysis was performed onpatients with a baseline Gastric Outlet Obstruction Score (GOOS) of0, 1, or 2 by evaluating the increase of GOOS by at least 1 point atany visit compared to baseline.

Seven patients with a baseline GOOS of 3 were not includedin primary endpoint analysis as their baseline GOOS could notincrease.

Five secondary endpoints were also assessed: technical successdefined as accurate placement and expansion of the stent; tim-ing of GOOS improvement assessed as time to GOOS improvementand persistence of GOOS improvement; symptoms of gastric outletobstruction; survival and need for re-treatments; complications.

2.3.5. Study measuresStudy measures were collected on standardized case report

forms as follows:At baseline: Patient demographics, tumour aetiology, stricture

location, presence of distant metastases, GOOS, symptoms of gas-tric outlet obstruction, prior treatment of the duodenal obstruction,presence of common bile duct stricture, and prior and/or concomi-tant chemotherapy or radiation.

Page 3

G. Costamagna et al. / Digestive and Liver Disease 44 (2012) 37– 43 39

F rs on

A nal Wd

tup

oom(

m

2

Foacpcsinv

m1q

ig. 1. (a) A long duodenal stricture due to pancreatic cancer (arrows) (the 2 markefter contrast injection patency of the proximal jejunum is documented. (b) A Duodeeployment shows stent patency.

At time of procedure: Treatment of a concomitant biliary stric-ure, length of stents and endoscope (forward or side viewing)sed, technical problems during stent insertion, occurrence of postrocedural complications and their treatment and outcome.

At all follow-up visits: GOOS, presence or absence of symptomsf gastric outlet obstruction, complications and their treatment andutcome, development of biliary obstruction after stenting, place-ent of biliary stent, and concomitant chemotherapy after stenting

documented at 15 day follow-up only).At study exit: Reason for end of follow-up, namely death, retreat-

ent, loss to follow-up, or completion of the 9 month follow-up.

.4. Statistical analysis

Summary statistics were computed for the treated population.or categorical measures at each visit, summary statistics consistf frequency and percent of responses in each category, as wells the exact Clopper–Pearson 95% confidence interval for the per-ent of responses. Unless otherwise noted, the denominator of aercentage is the number of subjects with non-missing values. Forontinuous measures at each visit, summary statistics consist ofample size, mean, median, standard deviation, minimum, max-mum, and 95% confidence interval for the mean, based on theormal distribution. The sample size is the number of non-missingalues, unless noted otherwise.

Time to event data was analysed using the Kaplan–Meierethod. Events analysed were death, a GOOS increase of at least

, its subsequent reversal, reaching a GOOS of 2–3, its subse-uent reversal, and duodenal obstruction. Event rates were based

the catheter are 7 cm apart). The stricture is passed with a catheter and guidewire.allFlex 9 cm in length is deployed. (c) Contrast injection immediately after WallFlex

on available follow-up data, with information censored beyond9 months after stent placement. The cumulative event rate at 9months and its standard error were computed. Median time toevent and its 95% confidence interval were determined.

McNemar’s test was used to assess symptom improvements ateach follow up time point as compared to baseline.

3. Results

3.1. Patients and baseline characteristics

Between August 2006 and November 2008, 206 patients wereprospectively enrolled in the registry. The treated populationincluded 202 patients (49% male, mean age 70.21 (SD 13.08) years).Four patients were excluded from the treated population, namely2 patients due to failure to pass the guidewire through the obstruc-tion, therefore, no stent was placed, and 2 patients due to totalabsence of data regarding the stent placement procedure or patientfollow up.

Two hundred and two patients were stented. Unless noted oth-erwise, all incident reports below are based on 202 patients. Nopatients had a history of duodenal stenting prior to participation inthe registry. The number of patients enrolled at each centre rangedfrom 3 to 64 with a median of 12.

Baseline tumour characteristics are summarized in Table 1. Clin-

ical condition of patients at baseline was poor, 20% of patients wereinoperable due to comorbidities or advanced age, 32% of patientswere unresectable, and 58% of patients had advanced disease. Base-line symptoms of obstruction vomiting, nausea and post prandial

Page 4

40 G. Costamagna et al. / Digestive and L

Table 1Baseline tumour characteristics.

Measure Number of patients

MalignancyPancreatic cancer 51.5% (104/202)Gastric cancer 18.3% (37/202)Duodenal carcinoma 9.0% (18/202)Choloangiocarcinoma 5.9% (12/202)Gallbladder cancer 5.9% (12/202)Ampullary carcinoma 1.0% (2/202)Othera 8.4% (17/202)

Presence of distant metastasesb

Only local cancer 49.5% (100/202)Liver metastasis 29.2% (59/202)Multiple metastasis 16.3% (33/202)Peritoneal carcinosis 14.9% (30/202)Lung metastasis 4.5% (9/202)

Previous chemotherapy and/orradiation

18.3% (37/202)

Chemotherapy after stentplacement

6.9% (14/202)

Previous biliary drainage 49% (99/202)

a Other types: 9 cases of colon cancer, 2 cases of breast cancer, 1 case of unspecifiedadenocarcinoma, and 1 each of: metastatic right kidney cancer, urothelial carcinomawith peritoneal carcinosis, metastatic endometrial cancer, recurrence of cancer oft

epiapp

3

ps

psdrs

TP

m

he gastric cardia, and stage IV squamous cell carcinoma of the cervix.b A patient may have more than one type of metastasis to different organs.

pigastric tenderness were experienced in 90%, 73% and 64% ofatients, respectively. Baseline GOOS was 0 in 55% of patients, 1

n 30% of patients, 2 in 11% of patients, and 3 in 4% of patients withn average baseline GOOS was 0.62 (SD 0.81). Mean weight losser patient before stenting was 10.1 (SD 6.1) kg over an averagere-stenting reporting period of 3.2 (SD 2.4) months.

.2. Stent placement and technical success

Procedural details are summarized in Table 2. A majority of stentlacement procedures were conducted using a side viewing endo-cope (60%). In total, 212 stents were placed.

Technical success was achieved in 98% (CI, 95%, 99%) of stentlacements. Four patients did not attain technical success due totent migration during deployment. Of these migrations, 3 wereistal and 1 was proximal to the stricture. All 4 technical failures

esulted in the successful placement of another WallFlex during theame procedure.

able 2rocedural data.

Measure Number of patients

Location of stricturea

Gastric antrum 14.9% (30/202)Bulb/upper duodenal genu 45.0% (91/202)2nd duodenal portion 30.7% (62/202)3rd duodenal portion 16.8% (34/202)Anastomotic recurrence 4.5% (9/202)

Type of endoscope usedSide viewing 60.4% (122/202)Forward viewing 39.6% (80/202)

Stent placement (per patient)One stent 95.0% (192/202)Two stents 5.0% (10/202)

Number of stents placed6 cm 37.7% (80/212)9 cm 52.8% (112/212)12 cm 9.4% (20/212)

a A patient may have a stricture extending through more than 1 duodenal seg-ent.

iver Disease 44 (2012) 37– 43

3.3. Biliary stenting

Biliary stenting was common in study patients. Approximatelyhalf of the patients required biliary drainage prior to duodenalstenting (99/202). During the duodenal stent placement procedure38/202 patients were concomitantly treated for biliary strictures byeither implanting (28) or by exchanging (10) a biliary stent. Forty-three biliary stents were placed in 38 patients, of these 20 weremetal uncovered stents, 18 were metal covered stents, and 5 wereplastic stents.

3.4. Clinical success (GOOS) and symptoms of gastric outletobstruction

Overall GOOS level improved from 14% of patients at baselinetolerating soft solids or low residue/normal diet to 84% at 15 days,86% at 30 days, 81% at 90 days, 79% at 180 days, and 70% at 270days. At 90 days post stent placement compared to baseline, symp-toms of post-prandial epigastric tenderness, nausea and vomitingimproved in 49%, 55%, and 70% of patients, respectively.

Clinical success, the primary endpoint, was achieved in 91% (CI,86%, 95%) of patients with a baseline GOOS of 0, 1 or 2. In 82% (CI,76%, 85%) of patients who achieved clinical success, there was noworsening of GOOS at any later time compared to baseline.

Symptoms of gastric outlet obstruction improved significantlycompared to baseline. Incidence of symptoms of gastric outletobstruction at each visit is reported in Table 3.

Seven patients were excluded from primary endpoint analysisbecause their baseline GOOS was 3 (ability to eat low residue ornormal diet) and therefore could not increase. Despite the near nor-mal GOOS values at baseline, these patients were in poor clinicalcondition. At baseline 100% of these patients were experiencing atleast one symptom of GOOS with confirmed malignant duodenalstrictures; therefore indicating them for stenting.

3.5. Timing and persistence of GOOS improvement

The Kaplan–Meier analyses of clinical success, time to increaseof GOOS by at least one point compared to baseline, and of timeto failure to maintain the GOOS increase by at least 1 point arepresented in Fig. 2. By day 5 (CI, 4, 6) after stent placement, 50%of patients had already experienced a GOOS increase of at least 1point, and this increase persisted for a median of 184 (CI, 109, 266)days.

For those patients reaching a GOOS of 2 or 3, the Kaplan–Meieranalyses of time to reaching the GOOS of 2 or 3 and of time to failureto maintain the GOOS of 2 or 3 is presented in Fig. 3. By day 5 (CI, 4,7) after stent placement, 50% of these patients had already reacheda GOOS of 2 or 3, and this GOOS level was maintained for a medianof 91 (CI, 87, 182) days.

3.6. Survival

One hundred and forty-five (72%) patients died prior to com-pletion of 9 months of follow-up. Death occurred at a median of94 days (CI, 79, 112). Cause of death was cancer progression in 141(97%) patients, not cancer related in 3 (2%) patients, and cause ofdeath was unknown in 1 (1%) patient who died in hospice care andfor which study participation was discontinued after the 3 monthvisit. Of the three deaths not related to cancer, at times expressedin days after the duodenal stenting procedure, 1 was due to sud-

den cardiac death on day 144, 1 was due to sepsis after gallbladdernecrosis due to biliary stenting on day 10, and 1 was due to pneu-monia on day 29. Five patients were lost to follow-up and survivaldata were not captured.

Page 5

G. Costamagna et al. / Digestive and Liver Disease 44 (2012) 37– 43 41

Table 3Gastric outlet obstruction symptoms over time intervals.a

Symptoms of obstruction Baseline (n = 202) 15-Day follow-up(n = 183)

1-Month follow-up(n = 161)

3-Month follow-up(n = 86)

6-Month follow-up(n = 43)

9-Month follow-up(n = 23)

Post-prandial epigastrictenderness

64.4% (130/202) 16.0% (29/181) 14.6% (23/158) 24.7% (21/85) 16.7% (7/42) 14.3% (3/21)

p-Value <.0001 <.0001 <.0001 <.0001 0.0005Nausea 73.3% (148/202) 19.6% (35/179) 20.9% (33/158) 23.8% (20/84) 26.2% (11/42) 22.7% (5/22)p-Value <.0001 <.0001 <.0001 <.0001 0.0039

(26/101

follo

3

coi

lootwiai

FGtb

Vomiting 90.1% (182/202) 15.3% (28/183) 16.1%p-Value <.0001 <.00

a The three symptoms of gastric outlet obstruction improved significantly at each

.7. Complications

A total of 43 complications were reported in 41 patients. Of theseomplications, 12 were minor complications requiring medicationr no intervention and 31 were major complications requiringntervention.

The most frequently reported complication was gastric out-et obstruction (GOO) due to ingrowth and/or overgrowth whichccurred in 12.4% (25/202) of patients. Ingrowth within 3 weeksf stent placement occurred in 5 patients possibly associated withhe previously reported cheese-cutter effect [13,14]. Later ingrowth

as reported after a mean of 105 days (median 67, range 27–320)

n 13 patients. Overgrowth with or without ingrowth occurred after mean of 172 days (median 168, range 42–263) in 7 patients. Stentngrowth and/or overgrowth were managed by the placement of a

ig. 2. (a) Time to achieve Gastric Outlet Obstruction Score (GOOS) increase ≥1, event ratOOS of 3 as this score could not increase. (b) Maintaining Gastric Outlet Obstruction Sco

ime required for a patient to decrease to their baseline GOOS or below baseline GOOS, easeline GOOS of 3.

61) 21.4% (18/84) 16.7% (7/42) 18.2% (4/22)<.0001 <.0001 0.0001

w-up visit when compared to baseline; all p-values < 0.05 using McNemar’s test.

second stent in 18 patients, by gastrojejunostomy in 2 patients (1patient choice and 1 treated in another hospital), by enteral tubefeeding in 1 patients, and by dilation and argon plasma ablation in1 patient. There was no further treatment in 3 patients.

Transient periprocedural symptoms on the day of duodenalstent placement were reported in 3.0% (7/202) of patients, namelyabdominal pain in 6 patients and vomiting in 1 patient. Symptomswere self limiting in 3 patients and managed with pain medicationin the other 4.

Bleeding occurred in 3.0% (6/202) of patients after a mean of 91days (median 36, range 22–200). Blood transfusion was required

in 3 patients, bleeding was self limiting in 2 patients requiring noadditional treatment and argon plasma coagulation was success-fully performed in 1 patient (in this case bleeding was probablyrelated to the stent).

e ± 1.96 SE, n = 195 patients treated, CI 95%. Does not include patients with baselinere (GOOS) increase of ≥1 is defined as once a GOOS of at least 1 was achieved, the

vent rate ± 1.96 SE, n = 178 patients treated, CI 95%. Does not include patients with

Page 6

42 G. Costamagna et al. / Digestive and Liver Disease 44 (2012) 37– 43

Fig. 3. (a) Time to achieve Gastric Outlet Obstruction Score (GOOS) 2–3, event rate ± 1.96 SE, n = 173 patients treated, CI 95%. Does not include patients with baseline GOOSo s oncel e patie

wimtofm

1w

w

c

4

chpot[egi

f 2 or 3. (b) Maintaining a Gastric Outlet Obstruction Score (GOOS) 2–3 is defined aess than 2–3, event rate ± 1.96 SE, n = 149 patients treated, CI 95%. Does not includ

Stent migration was reported in 1.5% (3/202) of patients within 1eek of stent placement. The resulting interventions were dilation

n 1 patient, surgical gastrojejunostomy in one patient, and place-ent of another stent in 1 patient. An attempt was made to re-stent

he patient who received gastrojejunostomy; however, intubationf the duodenum was not possible; therefore, the patient was sentor a salvage gastrojejunostomy. There were no late duodenal stent

igrations.Perforation was experienced in 0.5% (1/202) of patients on day

5. The perforation was likely related to stent induced ischemia andas managed by surgical suture and gastrojejunostomy.

Food impaction was reported in 0.5% (1/202) of patients. Thisas treated by endoscopic clearance of the stent.

Overall, gastrojejunostomy was required in 4 cases as a result ofomplications.

. Discussion

The patients treated as part of this series had similar baselineharacteristics as they were unresectable, poor surgical candidates,ad a shorter life expectancy, and/or had advanced disease. Inatients like these, SEMS placement for malignant gastroduodenalbstruction is the preferred method of treatment, as gastrojejunos-omy is associated with high rates of morbidity and mortality

7,15–17]. The findings of this series confirm the safety andffectiveness of duodenal stenting for the palliation of malignantastroduodenal obstruction. In most cases, patients experiencedncreased GOOS and significant symptom relief.

a GOOS 2–3 was achieved, the time required for that patient to decrease to a GOOSnts with baseline GOOS of 2 or 3.

Use of the WallFlex Duodenal Enteral Stent had technical andclinical success rates of 98% and 91%, respectively. The presentstudy enrolled patients from various countries around the worldsuggesting that duodenal stenting is a reproducible procedure andthe technology is available in a variety of settings [1–4].

Technical success of SEMS placement and gastrojejunostomyhas been reported to have similar results, with lower complica-tions related to stent insertion [7,11,15]. Nearly two thirds of stentswere inserted using a side-viewing duodenoscope, suggesting thatthis is the preferred scope for duodenal stenting [18]. The side-viewing duodenoscope provides better visualization of the lumenand duodenal stricture, additionally the elevator assists with theorientation of the catheter as well as gripping the guidewire anddelivery system [19].

With regard to clinical success, many patients can resume oralintake or have increased oral intake ability, within days after duo-denal stenting [20,21]. Maetani et al. [15] compared duodenal stentplacement to gastrojejunostomy and noted that most patients willresume oral intake on an average of 1 day after stenting in com-parison to 9 days after gastrojejunostomy. Additionally, Del Pianoet al. compared SEMS to gastrojejunostomy and found that clinicalsuccess was significantly higher in the SEMS group, 92% versus 56%(p = 0.0067) [7]. By day 5 after stent placement in both the Pies-man series [14] and the series reported here approximately 50%

of patients achieved a GOOS increase of at least 1 point. Similarlyby day 7 after stent placement 56% of patients reached a GOOS2 or 3 in the Piesman series, compared to 62% of patients in thisseries.

Page 7

and L

rccc

iwpgwmtigrjr

ch[taclsfp

dtt

tSd

wpam

st

CGNS

A

i

[

[

[

[

[

[

[

[

[

[

[

[

G. Costamagna et al. / Digestive

In our study a majority of centres experienced clinical successates ranging from 93% to 100%. A minority of centres experiencedlinical success rates ranging from 33% to 86%. Reasons for clini-al failure at these centres were directly related to patient clinicalondition.

Even though quality of life (QOL) was not objectively assessedn this study, it is self evident that duodenal stenting offers patients

ith advanced neoplastic disease a good source of palliation. Arior randomised study by Mehta et al. compared laparoscopicastrojejunostomy to duodenal stenting when palliating patientsith malignant gastroduodenal obstruction; 1 month post treat-ent SF-36 questionnaires were completed for both groups and

he mean physical health score was significantly higher in the stent-ng group when compared to the laparoscopic gastrojejunostomyroup 1 month after treatment (p < 0.01) [8]. Despite this result,e-treatment for stent obstruction can affect QOL and gastroje-unostomy is a possible alternative to stent insertion in patient inelatively good clinical condition.

The use of SEMS for gastroduodenal obstruction has been asso-iated with lesser procedural costs, shorter recovery time, earlierospital discharge, and less morbidity when compared to surgery5–10]. Duodenal stenting can be performed with fewer hospi-al resources as an operating room and anaesthesiologist are notlways necessary [7]. Additionally, procedural and hospitalisationosts for duodenal stenting combined have been shown to be muchess than that of surgery. Indeed, Yim et al. compared duodenaltenting to surgical gastrojejunostomy and found a significant dif-erence in median costs incurred (p < 0.005) as well as median postrocedural hospitalisation stay required (p < 0.005) [22].

Apart from short and long term complications related to stentysfunction due to tumour ingrowth or overgrowth, the complica-ion rate for this series was negligible, demonstrating once morehe safety of the duodenal stenting procedure.

A shortcoming of the registry was that the study was not con-rolled, namely not randomly compared to outcomes after surgery.uch randomised studies are, however, notoriously complex to con-uct.

In conclusion, adoption of duodenal stenting in hospitals world-ide along with adequate training can result in this mini-invasiverocedure becoming the standard of care to palliate debilitatednd frail patients experiencing GOO and suffering from advancedalignancy.This report confirms the safety and effectiveness of duodenal

tenting for palliation of malignant gastroduodenal obstruction inhe largest international prospective series to date.

onflict of interest. Costamagna, J. Spicak, A. Roy, T. Ponchon, H. Messmann, H.euhaus, and R. Laugier have a consulting agreement with Bostoncientific Corporation.

cknowledgements

This trial was sponsored by Boston Scientific Corporation (Nat-ck, Massachusetts, USA). The sponsor participated in planning

[

iver Disease 44 (2012) 37– 43 43

and executing the registry and in conducting data analysis. In thisregard, principal contributions were made by Montserrat AugustiPaloma, Eduardo Sessa, Alexis James, Joy Peetermans, PhD, medicalmonitor Robert S. Walsh, MD, and biostatisticians John Evans, PhD,Terry Liao PhD, Matthew Rousseau, Brian Johnson, and Wen Ding.

References

[1] Dormann A, Meisner S, Verin N, et al. Self-expanding metal stents for gas-troduodenal malignancies: systematic review of their clinical effectiveness.Endoscopy 2004;36:543–50.

[2] Telford J, Carr-Locke D, Baron T, et al. Palliation of patients with malignant gas-tric outlet obstruction with the enteral Wallstent: outcomes from a multicenterstudy. Gastrointest Endosc 2004;60:916–20.

[3] Maire F, Hammel P, Ponsot P, et al. Long-term outcome of biliary and duodenalstent in palliative treatment of patients with unresectable adenocarcinoma ofthe head of the pancreas. Am J Gastroenterol 2006;101:735–42.

[4] van Hooft J, Uitdehaag M, Bruno M, et al. Efficacy and safety of the newWallFlex enteral stent in palliative treatment of malignant gastric outletobstruction (DUOFLEX study): a prospective multicenter study. GastrointestEndosc 2009;69:1059–66.

[5] Maetani I, Ukita T, Tada T, et al. Metallic stents for gastric outlet obstruction:reintervention rate is lower with uncovered versus covered stents, despitesimilar outcomes. Gastrointest Endosc 2009;69:806–12.

[6] Mittal A, Windsor J, Woodfield J, et al. Matched study of three methods forpalliation of malignant pyloroduodenal obstruction. Br J Surg 2004;91:205–9.

[7] Del Piano M, Ballare M, Montino F, et al. Endoscopy or surgery for malignant GIoutlet obstruction? Gastrointest Endosc 2005;61:421–6.

[8] Mehta S, Hindmarsh A, Cheong E, et al. Prospective randomized trial oflaparoscopic gastrojejunostomy versus duodenal stenting for malignant gastricoutflow obstruction. Surg Endosc 2006;20:239–42.

[9] Maetani I, Tada T, Ukita T, et al. Comparison of dudoenal stent placement withsurgical gastrojejunostomy for palliation in patients with duodenal obstruc-tions caused by pancreaticobiliary malignancies. Endoscopy 2003;36:73–8.

10] Siddiqui A, Spechler S, Huerta S. Surgical bypass versus endoscopic stent-ing for malignant gastroduodenal obstruction: a decision analysis. Dig Dis Sci2007;52:276–81.

11] Jeurnink S, Steyerberg E, van Hooft J, et al. Surgical gastrojejunostomy orendoscopic stent placement for the palliation of malignant gastric outletobstruction (SUSTENT study): a multicenter randomized trial. GastrointestEndosc 2010;71:490–9.

12] Adler DG, Baron TH. Endoscopic palliation of malignant gastric outlet obstruc-tion using self-expanding metal stents: experience in 36 patients. Am JGastroenterol 2002;97:72–8.

13] van Hooft J, Mutignani M, Repici A, et al. First data on the palliative treatment ofpatients with malignant gastric outlet obstruction using the WallFlex enteralstent: a retrospective multicenter study. Endoscopy 2007;39:434–9.

14] Piesman M, Kozarek R, Brandabur J, et al. Improved oral intake after pallia-tive duodenal stenting for malignant obstruction: a prospective multi-centerclinical trial. Am J Gastroenterol 2009;104:2404–11.

15] Maetani I, Tada T, Ukita T, et al. Comparison of duodenal stent placement withsurgical gastrojejunostomy for palliation in patients with duodenal obstruc-tions caused by pancreaticobiliary malignancies. Endoscopy 2004;36:73–8.

16] Katsanos K, Sabharwal T, Adam A. Stenting of the upper gastrointestinal tract:current status. Cardiovasc Intervent Radiol 2010;33:690–705.

17] Brimhall B, Adler DG. Enteral stents for malignant gastric outlet obstruction.Gastrointest Endosc Clin N Am 2011;21:389–403.

18] Laasch HU, Martin DF, Maetani I. Enteral stents in the gastric outlet and duo-denum. Endoscopy 2005;37:74–81.

19] Boskoski I, Tringali A, Familiari P, et al. Self-expandable metallic stents formalignant gastric outlet obstruction. Adv Ther 2010;27:691–703.

20] Wong Y, Brams D, Munson L, et al. Gastric outlet obstruction secondary topancreatic cancer. Surg Endosc 2002;16:310–2.

21] Lindsay J, Andreyev H, Vlavianos P, et al. Self-expanding metal stents for the

palliation of malignant gastroduodenal obstruction in patients unsuitable forsurgical bypass. Aliment Pharmacol Ther 2004;19:901–5.

22] Yim HB, Jacobson BC, Saltzman JR, et al. Clinical outcome of the use of enteralstents for palliation of patients with malignant upper GI obstruction. Gastroin-test Endosc 2006;53:329–32.