BACKGROUND AND AIMS:

The aim of the study was to examine whether endoscopic intralesional corticosteroid injection into recalcitrant peptic esophageal strictures reduces the need for repeat stricture dilation.

METHODS:

Patients with a peptic esophageal stricture and recurrent dysphagia having had at least one dilation in the preceding 18 months were enrolled in a prospective randomized, double-blind study comparing steroid and sham injection. After endoscopic confirmation of recurrent stricture, patients were randomized to receive either 0.5 cc/quadrant triamcinolone (40 mg/cc) or sham injection into the stricture followed by balloon dilation of the stricture. Patients were stratified by the number of dilations required in the preceding 18 months, severity of dysphagia, the presence of esophagitis, stricture severity, and prior therapy with a proton-pump inhibitor. Patients and their physicians were blinded to the type of intervention received. Baseline dysphagia questionnaires were completed. Post-procedurally all patients were placed on a standardized proton-pump inhibitor regimen and standardized telephone follow-up questionnaires were completed at 1 wk and at 1, 3, 6, 9, and 12 months. The original sample-size calculation of 60 patients could not be met in a timely fashion because of a low incidence of recalcitrant peptic stricture patients.

RESULTS:

A total of 30 patients were enrolled, 15 in the steroid group (10 men, mean age 66 yr) and 15 in the sham group (11 M, mean age 67 yr). Patients were followed for 1 yr, unless they underwent an antireflux operation or died. Two patients, one per group, died of non-esophageal causes at 1 and 12 months. Four patients had fundoplication, two in each group, unrelated to stricture or dysphagia. Two patients in the steroid group (13%) and nine in the sham group (60%) required repeat dilation (p= 0.011).

CONCLUSIONS:

In patients with recalcitrant peptic esophageal stricture, steroid injection into the stricture combined with acid suppression significantly diminishes both the need for repeat dilation and the average time to repeat dilation compared to sham injection and acid suppression alone.

INTRODUCTION

An estimated 60–70% of benign strictures of the esophagus are peptic in origin and result from acid-induced mucosal ulceration (¹). The prevalence of strictures among gastroesophageal reflux disease (GERD) patients undergoing endoscopy is 4–20% (²). These individuals routinely undergo stricture dilation with standard polyvinyl tapered dilators or balloon dilators with a high initial rate of symptomatic improvement. However, 46–56% of patients undergoing dilation will require subsequent dilation within 1 yr (^3,4,5). Furthermore, two thirds of those patients requiring repeat dilation within the first year will require subsequent dilations (³). Potent acid suppression with proton-pump inhibitors may reduce the recurrence rate of strictures to 30% over a 1-yr period (⁶).

Various investigators have examined the role of corticosteroid injection into the stricture for prevention of recurrent esophageal strictures. Many of these studies have been small, uncontrolled investigations involving patients with strictures of diverse etiology (^{7,8,9,10,11,12}). Lee et al. retrospectively reported 12 non-blinded patients with refractory peptic esophageal strictures treated with four-quadrant injections of triamcinolone acetonide and standard dilation techniques (¹¹). There was a significant increase in the interval between dilations from 23 5 days to 167 49 days following treatment based on subjective symptoms (¹¹). Zein, Greseth, and Perrault demonstrated a similar reduction in the need for further stricture dilation following local steroid therapy in strictures of multiple etiologies (¹²). More recently, a randomized trial of Savary dilation with or without intralesional corticosteroids showed prolonged benefit in dysphagia and dilation frequency among patients receiving steroid injection (¹³). The study, however, did not control for the type of acid-suppressive therapy (H2-receptor antagonists vs proton-pump inhibitors) that has been shown in several studies to have variable influence on stricture recurrence rates (^6,14,15,16).

Some clinicians perform intralesional steroid injection for recurrent peptic esophageal strictures of the esophagus. Current data do not allow accurate conclusions to be made regarding the efficacy of this therapy. These studies have been limited by the lack of standardization of technique and blinding of investigators, inclusion of patients with both peptic and non-peptic strictures, and lack of standardization in symptom scoring and determinations of need for repeat dilation. A randomized, double-blinded, controlled investigation was undertaken to help resolve this question.

We hypothesized that intralesional injection of peptic esophageal strictures with a corticosteroid agent would reduce the requirement for repeat stricture dilation and improve dysphagia. Our primary aim was to compare differences in the need for repeat dilation 1 yr after intervention with dilation with versus without intralesional steroid injection.

METHODS

Protocol

The study was conducted as a randomized, double-blind, placebo-controlled investigation of two management strategies for patients with recurrent peptic esophageal stricture. Patients referred to our open access endoscopy unit for an esophagogastroduodenoscopy (EGD) for the indication of dysphagia were recruited in the study by an endoscopy nurse coordinator. Institutional Review Board approval was obtained and all participants gave written informed consent. Eligibility criteria included patients with dysphagia occurring at least once a week with a history of peptic stricture dilation within the preceding 18 months. When the findings of GERD such as erosions or ulcers of the distal esophagus were absent on endoscopy, patients were required to have had symptoms compatible with GERD as defined by validated criteria (¹⁷); heartburn at least once a week plus one associated symptom (relief with an antacid, nocturnal awakening, or radiation toward the neck) or acid regurgitation occurring at least once a week. Exclusion criteria included Barrett's esophagus with dysplasia, prior radiation therapy to the thoracic cavity, esophageal malignancy, prior esophageal or gastric resection, esophageal varices, or a clinical history implying that a stricture may be secondary to pill-induced esophagitis. Our report is compliant with the Consort statement on randomized trials (¹⁸).

Endoscopic injections were performed using a standard 22-gauge sclerotherapy needle (Hobbs, Stafford Springs, CT, USA). Triamcinolone acetonide (Kenalog^®, 40 mg/mL) was injected into the narrowest region of the stricture in the treatment group patients. Four-quadrant injections of 0.5 mL aliquots (20 mg) were performed prior to balloon dilation. Patients in both the study arms underwent similar dilation procedures with a standard through-the-scope balloon dilator (Microvasive MaxForce TTS dilator, Microvasive/Boston Scientific Corporation, Watertown, MA, USA). The initial balloon size was chosen based on the endoscopist's estimation of stricture diameter. The balloon was inflated with water to its recommended pressure for 30–60 s. The endoscopist was allowed to select further balloon diameters based on stricture characteristics, patient tolerance of the procedure, and the diameter required to relieve dysphagia. Strictures were dilated to a minimum diameter of 15 mm and a maximum diameter of 18 mm (^3,4,19). Patients in both treatment arms were prescribed esomeprazole 40 mg by mouth twice per day 30 min prior to morning and evening meals.

Assignment

A stratified randomization generated in the Division of Biostatistics was implemented by the Mayo Research Pharmacy (Saint Marys Hospital Rochester, MN, USA). The strata were specified by the number of dilations required in the preceding 18 months (1–2 vs 3 or more), severity of dysphagia (intermittent dysphagia for solids vs constant dysphagia for solids or dysphagia for liquids), the presence of esophagitis (normal mucosa vs linear erosions or ulceration), stricture severity (mild—endoscope passed without resistance vs moderate/severe—resistance to passage), and prior therapy with a proton-pump inhibitor. The above criteria were documented and relayed to the Mayo Research Pharmacy during the endoscopic examination. Patients were then randomized to active or sham injection within their specific strata. The block size within strata was two, and the assignments were generated based on a random number generator (function RANUNI) in the SAS^® software package (SAS Institute, Cary, NC, USA). Injection and dilation were performed as described above.

Masking

Patients and physicians performing the medical evaluation were blind as to the type of injection (steroid or sham) used by the endoscopist. The endoscopist was not involved in the clinical care of the patient. The contents of the syringe were concealed from the patient's view. The type of injection was recorded in the study records but was not noted in the patient's medical record or procedure note. Therefore, the type of injection an individual received was not identifiable during subsequent clinical evaluations. The individual who performed the follow-up telephone interviews and medical evaluations was unaware of the treatment arm. This negated potential bias in determining the need for subsequent dilation.

Participant Flow and Follow-Up

Patients were asked to rate the severity of their dysphagia on two scales adapted from Cox et al. (²⁰) and Marks et al. (¹⁴): dysphagia frequency (none, <1/wk, weekly, daily, each meal, each swallow, cannot eat) and dietary consistency intolerance (none, meat, bread, apple, banana, grits, water). Dysphagia frequency and dietary intolerances were evaluated at 1 wk, 1 month, 3 months, 6 months, and every 6 months thereafter, and at study completion as a mailed questionnaire. Standardized telephone interviews were conducted at 1 wk, 1 month, 3 months, 6 months, and every 6 months thereafter, and at study completion to detect persistent or worsening dysphagia. The need for repeat endoscopy with stricture dilation was based on the patient's report of dysphagia at least once a week during a telephone interview or a patient-initiated call, and confirmed during a face-to-face medical evaluation. Stricture dilation during subsequent procedures was similar to the initial procedure where patients received either steroid or sham injections based on their initial randomization. An overview of participant flow is represented in Figure 1.

Figure 1.

Study flow.

Full figure and legend (36K)

At each patient follow-up telephone call, pill counts were performed to assess compliance by asking "how often are you taking your esomeprazole 0%, 25%, 50%, 75%, or 100% of the time?" Patients who were unable to respond were asked to get their medication bottle and count aloud the number of pills remaining in their bottle while on the telephone. Telephone follow-up was required to secure delivery of complimentary proton-pump inhibitor therapy. Patients were instructed to avoid over-the-counter or prescription H2-receptor antagonists and prokinetic agents. Over-the-counter antacids were permitted as needed.

Patients were followed for a minimum of 1-yr post index dilation. The number of subsequent (repeat) dilation episodes was recorded and dysphagia scores were obtained via the questionnaires.

Sample Size

The primary endpoint in this study was the proportion free of repeat dilation over a 1-yr follow-up period, in each group. Assuming that 60 patients could be recruited over a 2-yr period with minimum follow-up of 1 yr on each patient, a sample size of 30 patients per group would have provided approximately 80% power to detect a difference in the proportions free of repeat dilation at 1-yr follow-up of 80% versus 50% or 90% versus 65%, based on a two-sample log-rank test at = 0.05. In the last part of the study period, the number of patients found to have peptic strictures significantly decreased. This change was likely due to the increasing use of proton-pump inhibitors in our clinical practice. It appeared that the study could not reach the targeted sample size in a timely fashion and a decision was made to terminate the study prematurely. The study recruited 15 patients per group over 2.5 yr and, with a minimum of 1-yr follow-up, had approximately 60% power to detect the above specified differences in the proportions free of repeat dilation at 1 yr (based on a two-sample log-rank test at = 0.05 assuming uniform accrual).

Analysis

A logistic regression analysis assessed the univariate association of both continuous and categorical baseline characteristics with group assignment. The primary analysis compared the time to repeat dilation during the follow-up period, which was designed to be a minimum of 1 yr per patient. A log-rank test was used to compare survival free of repeat dilation. Secondary analyses examined dysphagia frequency and food intolerance subcategories using Fisher's exact test.

RESULTS

Baseline patient characteristics are shown in Table 1. A total of 30 patients were enrolled (15 in each group) between 1999 and 2002. In the steroid group, there were 10 men and 5 women with a mean age of 66 yr. In the sham group, there were 11 men and 4 women with a mean age of 67 yr. The groups were also similar in regard to prior use of proton pump inhibitors (PPIs); non steroidal antiinflammatory drug (NSAID) use prior to enrollment and during the study; dysphagia frequency and severity; stricture length, location and severity; and presence of hiatal hernia, reflux esophagitis, and Barrett's esophagus at baseline endoscopy. The steroid group had higher body mass index (BMI) than the sham group (Table 1).

Table 1 - Baseline Patient Characteristics.

Full table

Patients were followed for a minimum of 1 yr unless death occurred or antireflux surgery was performed, at which point they were censored. Two patients, one per group, died of non-esophageal causes at 1 and 12 months. The patient who died at 1 month was censored; questionnaire data beyond baseline was not available. Four patients underwent fundoplication, two in each group, unrelated to stricture or dysphagia. Because two of the patients had surgery before the 1-yr date, these patients contributed observation time only up to the time of surgery. One patient provided 1-yr follow-up questionnaire data, then proceeded to surgery 3 months later for refractory GERD symptoms. Another patient, who eventually had an operation, had an event (i.e., repeat dilation) prior to 1-yr and had surgery 1 month later. Hence, overall follow-up was available in 30 of 30 (100%) subjects at the time of their first event (death, surgery, repeat dilation, or 1-yr), with 6-month questionnaire data available in 26 of 30 (87%) subjects and 1-yr questionnaire data available in 24 of 30 (80%) subjects. Compliance was roughly equivalent in both groups, with 92% of subjects in the steroid arm of the study and 94% of subjects in the sham arm of the study relaying compliance with medications more than 75% of the time.

Two of 15 (13%, 95% CI 4–38%) patients in the steroid group and 9 of 15 (60%, 95% CI 36–80%) in the sham group required repeat dilation (p= 0.0209). Figure 2 shows the time to first repeat dilation. Accounting for censoring, there was a shorter time to repeat dilation in the sham group (p= 0.01, log-rank test). Dysphagia frequency and dietary consistency intolerance at baseline and at time of first repeat dilation or 6 months (whichever occurred first) were compared between groups. Results are shown in Tables 2 and 3, respectively. No adverse events occurred either as a result of the dilation or steroid injection.

Figure 2.

Time to first repeat dilation.

Full figure and legend (13K)

Table 2 - Dysphagia Frequency.

Full table

Table 3 - Dietary Consistency Intolerances.

Full table

DISCUSSION

This study shows that intralesional corticosteroid injection of recalcitrant peptic esophageal strictures combined with acid suppression significantly diminishes both the need for repeat dilation and the average time to repeat dilation, compared to sham injection with acid suppression alone. It appeared that stricture recurrence was not only more likely to occur in the control group but also to occur sooner, typically in the first 6 months. We also noted that even with the prior use of proton-pump inhibitors in most patients, and their continued use during the study, there was a significant recurrence rate of strictures in the control group. BMI was the only baseline characteristic that differed significantly between groups, with subjects in the steroid treatment group having a higher BMI. The association between obesity and reflux is reasonably well accepted, with obese subjects being more prone to reflux and likely more prone to resistant strictures. Despite this association, subjects in the steroid injection arm of the study gained more benefit than their less obese counterparts who received sham injection.

Similarly, all subjects but one had short strictures, which would facilitate the endoscopists decision making regarding the location for injection. The single subject with a long stricture was randomized to receive intraluminal steroid injection. Theoretically, this subject would be most likely to fail due to the need to distribute the injections more widely. Despite this, patients randomized to intraluminal steroid injection demonstrated the most benefit.

The difference in dysphagia frequency and dietary consistency intolerance, shown in Tables 2 and 3, respectively, did not reach statistical significance at the time of follow-up. This is likely related to small sample size and the low event rate of recurrent dysphagia. With a larger study, we hypothesize that the trend toward favorable results with steroid injection would become statistically significant.

This prospective study confirms the findings of the previous retrospective case series. Kochhar and Makharia reported that 14 peptic stricture patients had a decreased frequency of repeat dilation in 6.5 months of follow-up with steroid injection (²¹). Similarly, the three peptic stricture patients in the series by Zein, Greseth, and Perrault did not require repeat dilation during a mean follow-up of 2 yr (¹²). Our study was unique in controlling acid therapy by providing proton-pump inhibitors throughout the study period. This was done to limit the role of acid-suppressive therapy in potentially confounding the results; a step not taken in previous reports (^{6,13,14,15,16}). Our study was also unique in that a relatively large standardized amount of corticosteroid was administered to each subject, rather than a discretionary dose, as was administered in other studies.

Triamcinolone's mechanism of action in preventing stricture recurrence in the esophagus is likely the same as the known corticosteroid effect on skin during wound healing. It has been shown that corticosteroids inhibit the inflammatory response to injury and decrease subsequent collagen formation. A single dose of methylprednisolone before a skin incision decreases wound strength by inhibiting angiogenesis, synthesis of granulation tissue, and wound contraction in rats (²²). Similar mechanisms are likely at work in the esophageal mucosa.

The strengths of this study include study design, randomization, close follow-up, and blinding of the patient, investigator, and primary physician. The magnitude of difference between the two groups was quite surprising and it is only due to the strengths in study design that we feel bias was limited as much as possible. Because neither the investigator nor the study coordinator initiated repeat EGD for dilation, these results are generalizable to clinical practice.

The main limitation of this study was failure to enroll the target number of participants because of the diminishing incidence of patients with peptic strictures seen in our practice. While this caused early termination of enrollment, it is likely that differences between groups would remain significant with a larger cohort.

It is possible that subjects who were non-compliant with daily proton-pump inhibitor therapy might experience more reflux and hence, have recurrent stricture formation more often than their compliant counterparts. Telephone pill counts and Research Pharmacy records (which mailed complimentary PPI to those who answered telephone follow-up) suggest that compliance was generally high among all participants and equal between the treatment groups. The literature shows that pill counts, whether performed in person or by telephone, and patient self-report, are flawed showing great variability and poor correlation with physiologic markers (^23,24,25). In other diseases, particularly hypertension, in which there is a disconnection between the long-term consequence of non-compliance and what the patient feels, compliance is a major issue. In hypertension, a new electronic method for monitoring frequency of opening a medication bottle has been advocated (²⁶). In our study of subjects with severe GERD with recalcitrant peptic strictures, non-compliance with complimentary PPI likely played a marginal role in the outcome.

Although this study serves to validate the use of steroids in the treatment of recalcitrant peptic esophageal strictures, it should be emphasized that this only applies to a small subset of patients with GERD. Steroid injection should not be considered routine practice at this time. Remaining questions include the utility of intralesional steroid injection in patients with other causes of esophageal strictures such as radiation, eosinophilic esophagitis, and photodynamic therapy. Benign luminal strictures, outside of the esophagus, may also be responsive to intralesional corticosteroid injection. Further trials comparing combined steroid injection and dilation, with dilation alone, for these other strictures may be worthwhile.

References

1. Spechler, SJ. AGA technical review on treatment of patients with dysphagia caused by benign disorders of the distal esophagus. Gastroenterology 1999;117: 233–254 (see comment). | Article | PubMed | ChemPort |
2. Ben Rejeb, M, Bouche, O, Zeitoun, P. Study of 47 consecutive patients with peptic esophageal stricture compared with 3880 cases of reflux esophagitis. Dig Dis Sci 1992;37: 733–736.
3. Patterson, DJ, Graham, DY, Smith, JL, et al. Natural history of benign esophageal stricture treated by dilatation. Gastroenterology 1983;85: 346–350.
4. Ogilvie, AL, Ferguson, R, Atkinson, M. Outlook with conservative treatment of peptic oesophageal stricture. Gut 1980;21: 23–25.
5. Hands, LJ, Papavramidis, S, Bishop, H, et al. The natural history of peptic oesophageal strictures treated by dilatation and antireflux therapy alone. Ann R Coll Surg Engl 1989;71: 306–309; (discussion 309–310).
6. Smith, PM, Kerr, GD, Cockel, R, et al. A comparison of omeprazole and ranitidine in the prevention of recurrence of benign esophageal stricture. Restore Investigator Group. Gastroenterology 1994;107: 1312–1318 (see comment). | PubMed | ChemPort |
7. Berenson, GA, Wyllie, R, Caulfield, M, et al. Intralesional steroids in the treatment of refractory esophageal strictures. J Pediatr Gastroenterol Nut 1994;18: 250–252.
8. Holder, TM, Ashcraft, KW, Leape, L. The treatment of patients with esophageal strictures by local steroid injections. J Pediatr Surg 1969;4: 646–653.
9. Kirsch, M, Blue, M, Desai, RK, et al. Intralesional steroid injections for peptic esophageal strictures. Gastrointest Endosc 1991;37: 180–182.
10. Kochhar, R, Ray, JD, Sriram, PV, et al. Intralesional steroids augment the effects of endoscopic dilation in corrosive esophageal strictures. Gastrointest Endosc 1999;49: 509–513. | Article | PubMed | ChemPort |
11. Lee, M, Kubik, CM, Polhamus, CD, et al. Preliminary experience with endoscopic intralesional steroid injection therapy for refractory upper gastrointestinal strictures. Gastrointest Endosc 1995;41: 598–601. | Article | PubMed | ChemPort |
12. Zein, NN, Greseth, JM, Perrault, J. Endoscopic intralesional steroid injections in the management of refractory esophageal strictures. Gastrointest Endosc 1995;41: 596–598. | Article | PubMed | ChemPort |
13. Dunne, D, Rupp, T, Rex, D, et al. Five year follow-up of prospective randomized trial of savory dilation with or without intralesional steroids for benign gastroesophageal reflux strictures. Gastroenterology 1999;116: A152.
14. Marks, RD, Richter, JE, Rizzo, J, et al. Omeprazole versus H2-receptor antagonists in treating patients with peptic stricture and esophagitis. Gastroenterology 1994;106: 907–915 (see comment). | PubMed | ChemPort |
15. Starlinger, M, Appel, WH, Schemper, M, et al. Long-term treatment of peptic esophageal stenosis with dilatation and cimetidine: Factors influencing clinical result. Eur Surg Res 1985;17: 207–214.
16. Farup, PG, Modalsli, B, Tholfsen, JK. Long-term treatment with 300 mg ranitidine once daily after dilatation of peptic oesophageal strictures. Scand J Gastroenterol 1992;27: 594–598.
17. Locke, GR, Talley, NJ, Weaver, AL, et al. A new questionnaire for gastroesophageal reflux disease. Mayo Clin Proc 1994;69: 539–547.
18. Moher, D, Schulz, K, Altman, D. The CONSORT statement: Revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA 2001;285: 1987–1991. | Article | PubMed | ISI | ChemPort |
19. Agnew, SR, Pandya, SP, Reynolds, RP, et al. Predictors for frequent esophageal dilations of benign peptic strictures. Dig Dis Sci 1996;41: 931–936. | Article | PubMed | ChemPort |
20. Cox, JG, Winter, RK, Maslin, SC, et al. Balloon or bougie for dilatation of benign oesophageal stricture? An interim report of a randomised controlled trial. Gut 1988;29: 1741–1747. | Article | PubMed | ChemPort |
21. Kochhar, R, Makharia, GK. Usefulness of intralesional triamcinolone in treatment of benign esophageal strictures. Gastrointest Endosc 2002;56: 829–834.
22. Rayner, TE, Cowin, AJ, Robertson, JG, et al. Mitogenic whey extract stimulates wound repair activity in vitro and promotes healing of rat incisional wounds. Am J Physiol Regul Integr Comp Physiol 2000;278: R1651–R1660.
23. Rudd, P, Byyny, RL, Zachary, V, et al. Pill count measures of compliance in a drug trial: Variability and suitability. Am J Hypertens 1988;1: 309–312.
24. Friedman, RH, Kazis, LE, Jette, A, et al. A telecommunications system for monitoring and counseling patients with hypertension. Impact on medication adherence and blood pressure control. Am J Hypertens 1996;9: 285–292.
25. Pullar, T, Kumar, S, Tindall, H, et al. Time to stop counting the tablets? Clin Pharmacol Ther 1989;46: 163–168. | PubMed | ChemPort |
26. Choo, PW, Rand, CS, Inui, TS, et al. Validation of patient reports, automated pharmacy records, and pill counts with electronic monitoring of adherence to antihypertensive therapy. Med Care 1999;37: 846–857. | Article | PubMed | ChemPort |

Acknowledgements

We would like to thank Teresa Zais for her help in data collection and Alicia Buda for her help in typing the manuscript. This work was supported by an AstraZeneca Investigator Initiated Award. The sponsor had no role in the design, execution, or analysis of the trial. Yvonne Romero was supported in part by a grant from the NIH DK 02956 and Joseph A. Murray was supported in part by a grant from the NIH DK 57892.