ssmuco

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2013;11:795–801

Endoscopic Corticosteroid Injections Do Not Reduce Dysphagia AfterEndoscopic Dilation Therapy in Patients With Benign EsophagogastricAnastomotic Strictures

MEIKE M. C. HIRDES,* JEANIN E. VAN HOOFT,‡ JAN J. KOORNSTRA,§ ROBIN TIMMER,� MAX LEENDERS,*RINSE K. WEERSMA,§ BAS L. A. M. WEUSTEN,� RICHARD VAN HILLEGERSBERG,¶

MARK I. VAN BERGE HENEGOUWEN,# JOHN T. M. PLUKKER,** RENEE WIEZER,‡‡ JAQUES G. H. M. BERGMAN,‡

FRANK P. VLEGGAAR,* PAUL FOCKENS,‡ and PETER D. SIERSEMA*

*Department of Gastroenterology and Hepatology and ¶Department of Surgery, University Medical Center Utrecht, Utrecht; ‡Department of Gastroenterology andHepatology and #Department of Surgery, Academic Medical Center, University of Amsterdam, Amsterdam; §Department of Gastroenterology and Hepatology and**Department of Surgery, University Medical Center Groningen, Groningen; and �Department of Gastroenterology and ‡‡Department of Surgery, St Antonius Hospital,Nieuwegein, The Netherlands

BACKGROUND & AIMS: Benign anastomotic strictures are often difficult to treat. We assessed the efficacy of addingcorticosteroid injections to endoscopic dilation therapy with Savary bougienage.

METHODS: In a multicenter, double-blind trial, 60 patients (mean age, 63 � 9 years; 78% male) with anuntreated cervical anastomotic stricture after esophagectomy with gastric tube reconstruc-tion and dysphagia for at least solid food were randomly assigned to groups given 4quadrant injections of 0.5 mL triamcinolone (40 mg/mL, n � 29) or saline (controls, n �31) into the stricture, followed by Savary dilation to 16 mm. Dysphagia, complications, andquality of life were assessed after 1 and 2 weeks and 1, 3, and 6 months. The primary endpoint was a dysphagia-free period of 6 months.

RESULTS: In the corticosteroid group, 45% of the patients remained dysphagia-free for 6 months,compared with 36% of controls (relative risk, 1.26; 95% confidence interval, 0.68 –2.36; P �.46). Median time to repeat dilation was 108 days (range, 15–180 days) in the corticosteroidgroup vs 42 days (range, 17–180 days) for controls (P � .11). A median number of 2dilations (range, 1–7) was performed in the corticosteroid group vs 3 dilations (range, 1–9)in controls (relative risk, 0.76; 95% confidence interval, 0.42–1.38; P � .36). Two majorintervention-related complications occurred, 1 submucosal laceration in the corticosteroidgroup and 1 hemorrhage in the control group. Four patients in the corticosteroid group,but none of the controls, developed Candida esophagitis (P � .03).

CONCLUSIONS: Corticosteroid injections do not provide a statistically significant decrease in frequency ofrepeat dilations or prolongation of the dysphagia-free period in patients with benignanastomotic esophagogastric strictures. Dutch Trial Registration Number 2236.

Keywords: Clinical Trial; Treatment; Swallowing; Endoscopy.

Dysphagia caused by a benign esophageal stricture cansignificantly impair quality of life.1 Benign esophageal

strictures can be caused by peptic, corrosive, or radiation injurybut can also occur at the site of the anastomosis after digestivesurgery.2 In a recently published review on benign esophagealtrictures treated with Savary bougienage from 1963–2006, ahift was observed from peptic to anastomotic strictures as the

ost common type.3 This can be explained by the widespreadse of proton pump inhibitor (PPI) therapy since its commer-ialization in 1989 in combination with the increasing numberf patients undergoing surgery for esophageal cancer.3,4

Approximately 40% of patients develop an anastomotic stric-ture after esophagectomy with gastric tube reconstruction.2,4 – 8

In most cases, anastomotic strictures can effectively be treated

within 3 endoscopic dilation sessions.8,9 It has been suggestedthat in patients with anastomotic esophageal strictures, endo-scopic incisional therapy with a needle-knife is more effectivethan standard dilation.10 Hordijk et al,9 however, demonstratedin another randomized controlled trial (RCT) with previouslyuntreated strictures that incisional therapy did not reduce the

Abbreviations used in this paper: CI, confidence interval; MDQ, MayoDysphagia Questionnaire; PPI, proton pump inhibitor; RCT, random-ized controlled trial; RR, relative risk.

© 2013 by the AGA Institute1542-3565/$36.00

http://dx.doi.org/10.1016/j.cgh.2013.01.016

odmrg

ta

r

ggccpntue

(owFt0

796 HIRDES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 7

number of endoscopies needed to become dysphagia-free incomparison with dilation.

An alternative endoscopic treatment that has been suggestedto improve clinical outcome is esophageal dilation combinedwith corticosteroid injections.11 Although current evidencefrom retrospective and prospective case series is debatable, ex-perts have advocated this treatment option as a good alternativeto standard dilation.11–13 One small, sham-controlled RCT dem-

nstrated that corticosteroid injections in combination withilation therapy were superior to dilation alone in the treat-ent of previously dilated, peptic strictures with redilation

ates of 13% in the corticosteroid group vs 60% in the controlroup (P � .01).14 The pathogenesis of anastomotic and peptic

strictures differs considerably, however. Peptic strictures de-velop as a result of inflammation and ulceration caused byreflux of gastric acid,13 whereas ischemia of the proximal part ofhe gastric tube is believed to be the main contributor tonastomotic stricture formation.4 Nonetheless, in both peptic

and anastomotic strictures the esophageal mucosa is chroni-cally inflamed. We therefore hypothesized that corticosteroidinjections combined with dilation in anastomotic stricturescould reduce the number of repeat dilations needed andlengthen the dysphagia-free period.

We conducted a multicenter, double-blind RCT to evaluatewhether Savary bougienage in combination with intralesionalcorticosteroid injections is more effective than bougienage withsaline injections in patients with benign anastomotic stricturesafter esophagectomy with gastric tube reconstruction.

MethodsThis multicenter study was performed at the Depart-

ments of Gastroenterology and Hepatology of the UniversityMedical Center Utrecht, Utrecht, the Academic Medical Center,Amsterdam, the University Medical Center Groningen, Gro-ningen, and the St Antonius Hospital in Nieuwegein, the Neth-erlands. The study protocol was reviewed and approved by theMedical Ethics Committees of all 4 participating centers andregistered at the Dutch Trial Registration (Trial number 2236).The study was designed and executed according to the guide-lines reported in the Consolidated Standards of ReportingTrials statement (2010).15 All authors had access to data andeviewed and approved the final manuscript.

PatientsPatients who, according to Ogilvie et al,16 had dyspha-

ia grade 2 or more after transhiatal or transthoracic esopha-ectomy with gastric conduit and cervical anastomosis with aonfirmed anastomotic stricture during endoscopy were in-luded. Exclusion criteria were a previous dilation or stentlacement (after esophagectomy), suspicion of recurrent malig-ancy, R1 or R2 resection, active anastomotic leak or perfora-ion, known duodenal or gastric ulcers, and poor candidates forpper endoscopy. Patients gave signed informed consent beforendoscopy was performed.

Randomization and Endoscopic ProcedureTwo sets of computer-generated, blocked (4 sheets in

each block) randomization sheets, with information on the typeof study medication to be injected, were put in consecutivelynumbered, sealed opaque envelopes by hospital personnel not

involved in the study. w

Before randomization, the presence and severity of the anas-tomotic stricture were determined by upper endoscopy. A mildstricture was defined as a stricture that could be passed by astandard diagnostic endoscope (GIF-H180 or GIF-Q165, Olym-pus, Tokyo, Japan; diameter 9.8 mm); severe strictures werethose that could not be passed by a standard diagnostic upperendoscope.

Randomization was stratified for center and stricture sever-ity and performed during endoscopy by study personnel notinvolved in the follow-up of patients. Patients and investigatorswere blinded to the type of injection fluid used by the endos-copist. Study medication was not masked, and endoscopistswere not blinded. Upper endoscopies were performed underconscious sedation by using midazolam or propofol by 2 seniorendoscopists per center who were not involved in participants’follow-up or clinical care. In case of mild strictures, dilation to16 mm was generally achieved in a single endoscopy; in case ofsevere strictures, 2 or more endoscopies within 1 week wereperformed. If dilation to 16 mm was considered safe, 0.5-mLaliquots of triamcinolone (Kenacort 40 mg/mL; Bristol-MyersSquibb BV, Woerden, the Netherlands) or normal saline (Cen-trafarm Services BV, Etten-Leur, the Netherlands) were injectedin 4 quadrants of the anastomotic stricture with a 4-mm-long,23-gauge needle (Boston Scientific Interject, Natick, MA),followed by stepwise dilation with Savary Gilliard bougies(Wilson-Cook Medical Inc, Winston-Salem, NC) over a stiffmetal guidewire to 16 mm. Savary dilation was performedinstead of balloon dilation because this was common practicein the participating centers for treatment of cervical strictures.We chose to perform injections before dilation to avoid the riskof inducing an esophageal perforation by injecting in an alreadylacerated esophageal wall.

Follow-up and Data CollectionAt baseline, data on age, sex, date and indication of

esophagectomy, perioperative blood loss, postoperative compli-cations such as anastomotic leakage, and administration ofneoadjuvant chemoradiotherapy, anticoagulants, and PPI ther-apy were assessed. During endoscopy, severity and length of thestricture, degree of gastric fluid retention, leakage of injectedfluid, technical problems, and complications were reported.After inclusion, patients were treated with 40 mg omeprazole(or equivalent alternative PPI) once daily (or an equivalent)during follow-up, if not already prescribed.

At baseline and 1 and 2 weeks and 1, 3, and 6 months afterinclusion, patients were interviewed by telephone by a blindedinvestigator who used standardized questions on dysphagia,quality of life, complications, PPI use, and satisfaction withendoscopic dilation. Dysphagia was assessed with a 4-pointscale (grade 0, ability to eat a normal diet; grade 1, ability to eatsome solid foods; grade 2, ability to eat semisolid food; grade3, ability to swallow liquids only; and grade 4, complete ob-struction)16 and the validated Mayo Dysphagia QuestionnaireMDQ).17 In addition, patients were asked to score dysphagia dailyn the 4-point scale during the first month after inclusion andeekly thereafter. Quality of life was measured by using the Shortorm 36 questionnaire. Patients’ satisfaction with the adminis-ered therapy was scored on a visual analogue scale, with a score ofbeing very unsatisfied and a score of 10 being extremely satisfied

ith the administered therapy.

fAsduuoapws

tf1pbAt

July 2013 CORTICOSTEROIDS FOR BENIGN ANASTOMOTIC STRICTURES 797

In case of recurrent dysphagia for liquids or solids, reportedduring a patient-initiated call or during a follow-up contact, arepeat endoscopy was scheduled within 2 weeks, depending onthe severity of recurrent dysphagia. A repeat dilation was per-formed when a stricture was visualized during endoscopy incombination with symptoms of dysphagia. A total of 3 endo-scopic dilations to 16 mm with intralesional injections of studymedication were allowed to be performed in the same studyarm during the 6-month study period. When dysphagia re-curred after 3 dilations with injections, endoscopic dilation wasperformed according to the preference of the endoscopist.

Study End PointsPrimary study end point was the number of dysphagia-

free patients without the need for repeat dilation(s) after6-month follow-up. For study purposes, dilation was defined asa dilation of 16 mm or more. In some cases, predilations wereindicated to safely reach 16 mm, but these were not counted aseffective dilations. Secondary study end points were time torepeat dilation, total number of endoscopic dilations, interven-tion-related (major) complications, dysphagia (measured by theMDQ), quality of life (measured by the Short Form 36), andpatients’ satisfaction with the administered therapy (scored ona visual analogue scale) after 6 months of follow-up.

StatisticsFor sample size calculations, we retrospectively reviewed

the proportion of patients with untreated anastomotic stric-tures that remained dysphagia-free for 6 months after standarddilation. Approximately 35% of patients were dysphagia-free for6 months after the first dilation in our hospital in 2008. Wehypothesized that adding corticosteroid injections was clini-cally relevant when a doubling of dysphagia-free patients (70%)was achieved after 6 months. To detect this clinically relevantincrease of 35% dysphagia-free patients with a two-sided � of0.05 and a power of 0.80, at least 27 patients per arm wereneeded. To compensate for a loss to follow-up of 10%, we aimedto include a total of 60 patients.

Normally distributed continuous variables were expressed asmean (� standard deviation) and non-normally distributed vari-ables as median (range). Categorical data are shown with percent-ages. A �2 test or Fisher exact test was used to analyze categoricalvariables, and a Student t test or Mann–Whitney U test was usedor continuous variables, depending on the normality conditions.

Kaplan–Meier analysis was used to determine the dysphagia-freeurvival per treatment arm (with 95% confidence interval [CI]). Theifference in survival curves between both groups was examined bysing a log-rank test. Negative binomial regression analysis wassed to estimate the relative risk (RR) for the difference in numberf dilatation sessions per treatment arm. Dysphagia, quality of life,nd patient satisfaction scores over time were analyzed with re-eated-measures analysis of variance. Stratified subgroup analysesere performed for patients with narrow strictures at initial pre-

entation. A P value � .05 was considered statistically significant.

ResultsBetween April 2009 and October 2011, 90 patients were

screened for eligibility. Fifteen patients did not have an anas-tomotic stricture during endoscopy, and 15 patients were ex-

cluded for other reasons (Supplementary Figure 1). Sixty pa-

tients were randomized to receive triamcinolone (n � 29) orsaline injections (n � 31) and were included in the intention-o-treat analysis. Between randomization and end of studyollow-up (February 29, 2012), 3 patients in the control arm and

patient in the triamcinolone arm died. One randomizedatient was excluded after endoscopic predilation complicatedy an esophageal laceration with high suspicion of perforation.fter exclusion of these patients, 55 patients were included in

he per-protocol analysis.

Patient CharacteristicsThere were no significant differences between both

groups at baseline (Table 1). All patients underwent transhiatal(33%) or transthoracic (67%) esophagectomy with hand-sewn,cervical anastomosis located between 16 and 24 cm from theincisors. Twenty-one percent of patients developed an anasto-motic leak after surgery. Median time to stricture developmentafter surgery was 69 days (range, 0 –313), and median stricturediameter at initial presentation was 9 mm (range, 3–14). Duringfollow-up 29 of 31 and 28 of 29 patients in both groupscontinued their PPI use until end of follow-up or death.

Clinical EfficacyAfter 6 months of follow-up, 45% of patients (13 of 29)

who received steroid injections remained dysphagia-free vs 36%of patients (11 of 31) who were injected with saline (RR, 1.26;

Table 1. Patient Characteristics at Baseline

Triamcinolone(n � 29) (%)

Saline(n � 31) (%)

Pvalue

Mean age � standarddeviation (y)

64 � 9 62 � 8 .16

Male sex 25 (86) 22 (71) .15Indication for esophagectomy .51

Adenocarcinoma 22 (76) 23 (74)Squamous cell carcinoma 5 (17) 7 (23)Nonmalignant disorder 2 (7) 1 (3)

Perioperative blood loss (mL) .47Unknown 14 (48) 17 (55)None — 2 (7)�300 5 (17) 4 (13)�300 10 (35) 8 (26)

Anastomotic leak 4 (14) 9 (29) .15Median time to stricture

(range) (d)64 (34–313) 70 (0–216) .61

Stricture diameter (mm) .13�5 4 (14) 5 (16)5–10 16 (55) 15 (48)�10 9 (31) 11 (36)

Dysphagia score 5 (17) 6 (19) .64II 15 (52) 15 (49)III 9 (31) 10 (32)IV

Median stricture length(range) (mm)

4 (2–10) 4 (2–20) .82

Use of relevant medicationPPI 13 (45) 15 (48) .52Anticoagulant/antiplatelet

drugs14 (48) 12 (39) .46

Neoadjuvant chemotherapy 22 (76) 21 (68) .90

Adjuvant chemotherapy 3 (10) 3 (10) .93

ss

dpsl

de

tosttmtr

798 HIRDES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 7

95% CI, 0.68 –2.36; P � .46; Figure 1). Median time to repeatdilation was 108 days (range, 15–180) in the steroid group vs 42days (range, 17–180) in the saline group (log-rank test, P � .29).The median number of dilations performed after 6 months was2 (range, 1–7) in the steroid group vs 3 (range, 1–9) in the salinegroup (RR, 0.76; 95% CI, 0.42–1.37; P � .36). Per-protocolanalysis was also nonsignificant: 13 of 27 dysphagia-free pa-tients in the steroid group vs 9 of 28 in the saline group (RR,1.50; 95% CI, 0.77–2.91) and a median of 2 (1–7) vs 3 (1–9)dilations needed in 6 months of follow-up.

Subgroup analyses performed in patients with narrow stric-tures at initial presentation resulted in 31% dysphagia-free pa-tients (5 of 16) in the steroid group vs 13% in the saline group(2 of 16) (RR, 2.51; 95% CI, 0.61–11.13; P � .20). Median timeto repeat dilation was 40 days (95% CI, 2.76 –77.24) in thesteroid group vs 27 days (95% CI, 25.07–28.30) in the salinegroup (log-rank test, P � .07). The median number of dilationsperformed after 6 months was 2 (range, 1–7) in the steroidgroup vs 4 (range, 1–9) in the saline group (RR, 0.67; 95% CI,0.67–1.46; P � .31).

Dysphagia (measured by using the MDQ questionnaire) andquality of life gradually improved during follow-up, but nodifferences were observed between patients treated with steroidsor saline (Figure 2A and B). Overall patient satisfaction washigh but not different between groups (Figure 2C).

ComplicationsTwo major adverse events (3%) that occurred during

follow-up were related to endoscopic dilation (Table 2). Atinitial predilation, one patient allocated to corticosteroid injec-tions developed an esophageal laceration that was suspected tobe a perforation and resulted in exclusion from study partici-pation. He was treated conservatively with nil per mouth and

Figure 1. Kaplan–Meier analysis showing proportion of dysphagia-free patients during 6-month follow-up (log-rank test, P � .29).

antibiotics and was discharged uneventfully. During a repeat a

endoscopy 2 weeks later, the laceration had healed. One patientin the saline group developed melena and a hemoglobin dropafter endoscopic dilation. After the procedure, we learned thatshe had continued her oral anticoagulant use despite instruc-tions to discontinue these shortly before endoscopy. During arelook endoscopy a clot was visible at the location of theanastomosis. She was treated with 2 units of packed red bloodcells and recovered uneventfully. In the corticosteroid group, 4patients developed Candida esophagitis proximal of the anasto-mosis vs none in the saline group (P � .03). Patients treatedwith adjuvant chemotherapy had a statistically significanthigher risk of Candida esophagitis when compared with patientsnot treated with adjuvant therapy (2 of 6 patients on adjuvanttherapy vs 2 of 54 patients not on adjuvant therapy; RR, 9.0;95% CI, 1.53–52.79; P � .01).

DiscussionThis multicenter, double-blind RCT evaluated the effi-

cacy of adding intralesional steroid injections to endoscopicdilation in patients with anastomotic strictures after esopha-gectomy. We demonstrated that corticosteroid injections donot provide a statistically significant decrease in frequency ofrepeat dilations or a statistically significant prolongation of thedysphagia-free period. Furthermore, corticosteroid injectionsseem to be associated with an increased risk of Candida esoph-agitis, particularly in patients on adjuvant chemotherapy.

The effects of intraesophageal corticosteroid injections havebeen proposed to increase the efficacy of endoscopic dilationfor benign strictures since the early 1970s.11,13 Thus far, RCTsupporting this hypothesis are limited and, if available, are onlymall-sized and not focused on anastomotic strictures.14,18,19 A

Portuguese RCT did not demonstrate a difference in dilationfrequency or dysphagia in 14 patients with corrosive stricturesallocated to steroid injections or placebo. They only found adifference in the maximum achieved diameter of dilation.19 Inanother RCT, 21 patients with strictures of various etiologiesand variable numbers of previous dilations were included.18

These authors found an increase in dysphagia-free period andperiodic dilation index in the steroid arm but no difference intotal number of dilations. Another study, only published inabstract form, demonstrated a decrease in mean dilation fre-quency in patients with peptic strictures from 6 dilations in thecontrol group to 2 dilations in the steroid group after 1-yearfollow-up.20 Finally, Ramage et al14 demonstrated a significant

ifference in repeat dilations after 1 year of follow-up in 30atients with previously dilated peptic strictures (13% in theteroid group vs 60% in the placebo group) and a significantlyonger dysphagia-free survival.

How can the lack of convincing efficacy of adding steroids toilation in anastomotic strictures in our study be explained,specially in comparison with the above-mentioned studies?

The pathogenesis of anastomotic strictures is different fromhat of peptic or corrosive strictures, which develop as a resultf direct contact of the esophageal mucosa with chemical sub-tances, ie, gastric content or corrosive substances. During gas-ric tube reconstruction, the left epiploic artery is sacrificed, andhe blood supply of the proximal gastric conduit depends on

icroperfusion and collateral flow, which makes it susceptibleo ischemia. Although the exact mechanism is not clear, scar-ing and fibrosis resulting in stricture formation are thought to

rise from ischemia-induced chronic inflammation. This differ-

cbesas

M

July 2013 CORTICOSTEROIDS FOR BENIGN ANASTOMOTIC STRICTURES 799

ence in pathogenesis may well explain the disappointing re-sponse to corticosteroid injections in anastomotic strictures.

It can also be imagined that efficacy improves when cortico-steroid injections are used at an earlier stage, preferably beforestricturing has occurred. Corticosteroids are known to inhibitmigration and activation of polymorphonuclear leukocytes andfibroblasts and inhibit edema and collagen formation.21 Whenorticosteroids are used in the active phase of inflammationefore fibroblast proliferation has occurred, it may be moreffective than when collagen formation is already present. A recenttudy evaluated prophylactic corticosteroid injections in patientsfter semicircumferential esophageal endoscopic submucosal dis-ection.22 Although this was a single arm study, it did show a

convincing reduction of stricture formation from 75% to 19% (P �.01) when compared with a historical cohort.

Strengths of this study include the randomized study designwith stratification for stricture severity, blinding of patients andinvestigators, and the homogeneous study population, includ-ing only patients with untreated, hand-sewn, cervical anasto-motic strictures after esophagectomy. Furthermore, in the lightof the somewhat disappointing clinical efficacy, the relativelyhigh-dose triamcinolone chosen in this study cannot be heldaccountable for the lack of efficacy. On the other hand, it maywell explain the high incidence of symptomatic Candida esoph-agitis in patients in the corticosteroid group (14%).

The main limitations of this study are the relatively smallsample size and limited follow-up period. One may speculatethat our study was undersized to detect small differences be-tween the groups. We calculated that at least 200 –750 patientshad to be included to detect a significant improvement ofdysphagia-free patients of 10%–20% by using a two-sided � of0.05 and a power of 0.80. Besides the fact that including sucha sample size was not considered feasible, we also did not

4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™Figure 2. (A) Dysphagia scores measured with the MDQ. Scores�40% are considered positive for dysphagia, 15%–40% indetermi-nate, and �15% are negative for dysphagia. (B) Quality of life (QoL)scores measured with the Short Form 36. A score of 100 representsbest imaginable health, and a score of 0 represents worst imaginablehealth. (C) Patient satisfaction with endoscopic dilation therapy scored

Table 2. Complications

Triamcinolone(n � 29) (%)

Saline(n � 31) (%)

Pvalue

Major adverse events 6 (24) 8 (26) .86Upper gastrointestinal

bleedinga— 1

Deep esophageal lacerationa 1 —Food impaction 1 1Malignant recurrence 1 3Myocardial/cerebral infarct 1 2Dehydration due to diarrhea/

vomiting2 1

Respiratory complications — —inor adverse events 4 (14) 0 (0) .03Candida esophagitis 4 0

aMajor complications with probable association to study intervention.

on a visual analogue scale (0–10).

ptimols

hta

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

800 HIRDES ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 7

consider an increase in efficacy of 10% to be clinically relevant.Especially when taking the increased risk of Candida esophagitisinto account, which also induces odynophagia and dysphagia,we believed that the benefits of the additional interventionshould outweigh the limitations. Nonetheless, we must keep inmind that the current study did not have sufficient power todetect smaller clinical differences.

What are the implications of these study results for thetreatment of benign anastomotic esophageal strictures in dailypractice? Standard endoscopic dilation remains the treatmentof first choice, because its efficacy and safety are satisfactory inthe majority of patients.23,24 Electrocautery incisional therapyhas not been shown to reduce the number of dilations inpatients with anastomotic strictures.9 Our study also did not

rove superiority of adding corticosteroid injections to dilationherapy in the treatment of anastomotic strictures. Self-expand-ng stents remain effective continuous dilators but are also

uch more invasive than dilation therapy. Several recent seriesn stent placement in benign esophageal strictures have shown

imited clinical efficacy and often considerable complications,uch as retrosternal pain and stent migration.25–28 Whether the

benefit of prolonged dilation outweighs the relatively highcomplication risk associated with stent placement remains tobe determined. We are currently performing a multicenter RCTthat compares biodegradable stent placement with standarddilation therapy in patients with previously dilated benignesophageal strictures to provide an answer to this question.

In conclusion, this study shows that adding corticosteroidinjections to endoscopic dilation did not result in a statisticallysignificant reduction of repeat dilations and prolongation ofdysphagia-free interval. Moreover, intralesional corticosteroidinjections were associated with a statistically significant in-creased risk of Candida esophagitis. Although our study may

ave had insufficient power to detect minor benefits, the rou-ine use of corticosteroid injections in patients with benignnastomotic strictures cannot be recommended.

Supplementary MaterialNote: To access the supplementary material accompa-

nying this article, visit the online version of Clinical Gastroen-terology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2013.01.016.

References

1. Shah JN. Benign refractory esophageal strictures: widening theendoscopist’s role. Gastrointest Endosc 2006;63:164–167.

2. Lew RJ, Kochman ML. A review of endoscopic methods of esoph-ageal dilation. J Clin Gastroenterol 2002;35:117–126.

3. Piotet E, Escher A, Monnier P. Esophageal and pharyngeal stric-tures: report on 1,862 endoscopic dilatations using the Savary-Gilliard technique. Eur Arch Otorhinolaryngol 2008;265:357–364.

4. Honkoop P, Siersema PD, Tilanus HW, et al. Benign anastomoticstrictures after transhiatal esophagectomy and cervical esopha-gogastrostomy: risk factors and management. J Thorac Cardio-vasc Surg 1996;111:1141–1146.

5. Sutcliffe RP, Forshaw MJ, Tandon R, et al. Anastomotic stricturesand delayed gastric emptying after esophagectomy: incidence, riskfactors and management. Dis Esophagus 2008;21:712–717.

6. Williams VA, Watson TJ, Zhovtis S, et al. Endoscopic and symp-tomatic assessment of anastomotic strictures following esopha-gectomy and cervical esophagogastrostomy. Surg Endosc 2008;

22:1470–1476.

7. Heitmiller RF, Fischer A, Liddicoat JR. Cervical esophagogastricanastomosis: results following esophagectomy for carcinoma.Dis Esophagus 1999;12:264–269.

8. van Heijl M, Gooszen JA, Fockens P, et al. Risk factors fordevelopment of benign cervical strictures after esophagectomy.Ann Surg 2010;251:1064–1069.

9. Hordijk ML, van Hooft JE, Hansen BE, et al. A randomized com-parison of electrocautery incision with Savary bougienage forrelief of anastomotic gastroesophageal strictures. GastrointestEndosc 2009;70:849–855.

0. Hordijk ML, Siersema PD, Tilanus HW, et al. Electrocautery ther-apy for refractory anastomotic strictures of the esophagus. Gas-trointest Endosc 2006;63:157–163.

1. Ashcraft KW, Holder TM. The experimental treatment of esopha-geal strictures by intralesional steroid injections. J Thorac Car-diovasc Surg 1969;58:685–691.

2. Kochhar R, Ray JD, Sriram PV, et al. Intralesional steroids aug-ment the effects of endoscopic dilation in corrosive esophagealstrictures. Gastrointest Endosc 1999;49:509–513.

3. Spechler SJ. AGA technical review on treatment of patients withdysphagia caused by benign disorders of the distal esophagus.Gastroenterology 1999;117:233–254.

4. Ramage JI Jr, Rumalla A, Baron TH, et al. A prospective, random-ized, double-blind, placebo-controlled trial of endoscopic steroidinjection therapy for recalcitrant esophageal peptic strictures.Am J Gastroenterol 2005;100:2419–2425.

5. Schulz KF, Altman DG, Moher D, et al. CONSORT 2012 state-ment: updated guidelines for reporting parallel group randomisedtrials. BMJ 2010;340:c332.

6. Ogilvie AL, Dronfield MW, Ferguson R, et al. Palliative intubationof oesophagogastric neoplasms at fibreoptic endoscopy. Gut1982;23:1060–1067.

7. Grudell AB, Alexander JA, Enders FB, et al. Validation of the MayoDysphagia Questionnaire. Dis Esophagus 2007;20:202–205.

8. Altintas E, Kacar S, Tunc B, et al. Intralesional steroid injection inbenign esophageal strictures resistant to bougie dilation. J Gas-troenterol Hepatol 2004;19:1388–1391.

9. Camargo MA, Lopes LR, Grangeia TA, et al. [Use of corticoste-roids after esophageal dilations on patients with corrosive ste-nosis: prospective, randomized and double-blind study]. Rev As-soc Med Bras 2003;49:286–292.

0. Dunne D, Rupp T, Rex D, et al. Five year follow-up of prospectiverandomized trial of savory dilation with or without intralesionalsteroids for benign gastroesophageal reflux stricture. Gastroen-terology 1999;116:A152.

1. Miyashita M, Onda M, Okawa K, et al. Endoscopic dexamethasoneinjection following balloon dilatation of anastomotic stricture afteresophagogastrostomy. Am J Surg 1997;174:442–444.

2. Hashimoto S, Kobayashi M, Takeuchi M, et al. The efficacy ofendoscopic triamcinolone injection for the prevention of esopha-geal stricture after endoscopic submucosal dissection. Gastroin-test Endosc 2011;74:1389–1393.

3. Scolapio JS, Pasha TM, Gostout CJ, et al. A randomized prospec-tive study comparing rigid to balloon dilators for benign esopha-geal strictures and rings. Gastrointest Endosc 1999;50:13–17.

4. Pereira-Lima JC, Ramires RP, Zamin I Jr, et al. Endoscopic dila-tion of benign esophageal strictures: report on 1043 procedures.Am J Gastroenterol 1999;94:1497–1501.

5. van Boeckel PG, Vleggaar FP, Siersema PD. A comparison oftemporary self-expanding plastic and biodegradable stents forrefractory benign esophageal strictures. Clin Gastroenterol Hepa-tol 2011;9:653–659.

6. Hirdes MM, Siersema PD, Vleggaar FP. A new fully covered metalstent for the treatment of benign and malignant dysphagia: aprospective follow-up study. Gastrointest Endosc 2012;75:712–

718.

2

July 2013 CORTICOSTEROIDS FOR BENIGN ANASTOMOTIC STRICTURES 801

27. Eloubeidi MA, Talreja JP, Lopes TL, et al. Success andcomplications associated with placement of fully covered re-movable self-expandable metal stents for benign esophagealdiseases (with videos). Gastrointest Endosc 2011;73:673–681.

8. Bakken JC, Wong Kee Song LM, de Groen PC, et al. Use of afully covered self-expandable metal stent for the treatment ofbenign esophageal diseases. Gastrointest Endosc 2010;72:

712–720.

Reprint requestsAddress requests for reprints to: Meike M. C. Hirdes, MD, Depart-

ment of Gastroenterology and Hepatology, University Medical CenterUtrecht, P.O. Box 85500, 3508 GA Utrecht, Netherlands. e-mail: m.m.c.hirdes@umcutrecht.nl; fax: 31-88-755-5533.

Conflicts of interest

The authors disclose no conflicts.

Consolidated Standards of Reporting Trials (CONSORT) statement 2010.

July 2013 CORTICOSTEROIDS FOR BENIGN ANASTOMOTIC STRICTURES 801.e1

Supplementary Figure 1. Flow diagram of the study according to