Pereira Lima JC, Marroni CA, Lopes CV, Zamin Jr. I, Cassal AP
Porto Alegre, Brazil.

Background/aims: Endoscopic dilation is considered the treatment of choice for most cases of benign esophageal stricture, although it remains controversial if fluoroscopy is necessary and which kind of stricture is the most amenable to treatment. We report here on our experience on a prospectively collected series of 1310 dilation sittings without the aid of fluoroscopy and compare the results of this therapy among patients with strictures of different etiologies.

Methods: Between 5/1992 and 10/1998, we performed 1310 dilation sessions on 167 patients (median=5, range 1-56). Treatment was considered optimal if esophageal lumen could be dilated up to 42-45 French and there was no dysphagia after therapy. If the stricture was dilated up to 42-45 French and there was dysphagia to some solid foods, the response to treatment was judged good.

Results: 167 patients (111 M, 56 F, mean age=55.3 yr.) were followed-up for a mean of 18.8 (1-75) months. Stricture´s etiology was post-surgical in 96, peptic in 43, caustic in 12 and from other etiologies in 16. Optimal or good dilation was achieved in 146 of 167 patients (87.4%). Optimal response was achived in 60.5%, 69% and 36.4% of the patients with postoperative, peptic and caustic strictures, respectively (p=0.021). Optimal response was accomplished in 84% of the patients who presented with dysphagia to solids and in 41.5% of the cases with dysphagia to liquids (p=0.0001).

Conclusion: Endoscopic dilation without the aid of fluoroscopy is safe and effective in relieving dysphagia caused by benign strictures of different etiologies. Patients with severe dysphagia needed also more sessions to adequate dilation be achieved (3 vs. 7 sessions, p=0.001). There were 5 esophageal perforations, with 1 death (0.38% and 0.08% per session and 3% and 0.6% per patient, respectively). Patients with caustic stricture respond worse to this therapy than patients with postoperative or peptic stenoses. The dysphagia level at clinical presentation is associated with the number of sessions needed and with the success of treatment.

Key Words: esophageal strictures, benign peptic strictures, bougienage, esophageal dilation, endoscopy.


Endoscopic peroral dilation is considered the initial treatment approach of choice for patients with benign esophageal strictures (1, 2). The endoscopic dilation procedures can be performed with mercury-weighted rubber blunt-tipped ( Hurst type) or tapered-tipped ( Maloney type) bougies, wire-guided polyvinyl bougies ( Savary-Gilliard type), wire-guided metal olives (Eder-Puestow type) or endoscopically oriented balloons ( 3, 4 ).

Although endoscopic esophageal striture dilation is the chief cause of the most dreaded complication of a diagnostic or therapeutic upper digestive endoscopy, i.e. perforation, this technique remains considered the safest, and perhaps the most efficient method to mitigate both dysphagia and malnutrition caused by benign esophageal stenoses (5, 6). In this study, we report our experience on a prospectively assessed series of 153 patients with 1043 dilation sittings.



A total of 153 patients with benign esophageal stenosis seen consecutively between May 1992 and July 1997 were enrolled in the study. All patients presented with dysphagia of varying grades and the benign etiology of the stricture was confirmed by means of endoscopy with biopsies and by the clinical outcome.

Endoscopic Dilation

The dilation sessions were done as an outpatient procedure in the vast majority of the patients. A barium esophagogram was performed before the first dilation sitting only when the endoscopist suspected the existence of a tortuous stenosis ( e.g., all cases of caustic and some cases of peptic strictures). Dilation was done using Savary-Gilliard dilators (Wilson Cook Medical Inc., Winstom-Salem, USA) or Eder-Puestow dilators (Key Medical Inc., Essex, UK). Bougies or metal olives of increasing size were passed over a guide wire, which had been positioned with the help of an endoscope (GIFXQ20, CV-1 Video Endoscope or CV-100 Video Endoscope, Olympus, Corp., Tokyo, Japan). Usually, one to four dilators were passed during each session, depending upon patients tolerance and stricture tightness. Fluoroscopic control of the procedure was never performed. Dilation sessions were done on a weekly basis until a lumen size of 42 to 45 French ( 14 to 15 mm) was obtained. Afterwards, dilation was repeated whenever dysphagia recurred. Patients who had strictures caused by peptic esophagitis were advised to use proton pump inhibitors, such as omeprazole or lansoprazole, together with other antireflux measures such as raising the head of the bed.

After each dilation session, patients were carefully observed for any complication such as perforation or bleeding. If perforation was suspected, an urgent chest X ray and an esophagogram using a water soluble contrast medium were done. If perforation was confirmed, an immediate surgical consultation was sought and the patient was put on intravenous fluids, parenteral antibiotics and was advised to take nothing by the oral route.

Response to treatment

Response to treatment was judged on the basis of improvement in dysphagia, which was graded as follows: 0 – taking a normal diet, no dysphagia; 1 – unable to swallow certain solid foods ( e.g. grilled beef); 2 – able to swallow only semisolid soft foods; 3 – able to swallow just liquids; 4 – unable to swallow liquids in adequate amounts.

The response to endoscopic management was considered good if the stricture could be dilated up to a diameter of 42 to 45 French with complete or almost complete relief of dysphagia ( grade 0 and 1, respectively). When the stenosis could not be adequately dilated until this bougie diameter, or there was minimal or no improvement in the symptoms even if the stricture was dilated up to 45 French, the response to treatment was considered "poor"(grades 2, 3 and 4).

Treatment follow-up

After the initial adequate dilation, patients were instructed to return to our Endoscopy Unit for a control endoscopic examination one month and three months after the last endoscopic procedure or whenever dysphagia recurred. One year after the last dilation session, patients were contacted to assess response to treatment.


Kruskal-Wallis test was applied for verifying the significance of correlations between etiology and the number of sessions needed to achieve adequate response and dysphagia grade versus number of sessions. Chi-square test was employed to assess correlation between etiology and pre-treatment dysphagia grade and etiology versus response to treatment. Differences were considered significant only when a p value of < 0.05 was achieved.


We prospectively studied 153 patients, in whom we performed 1043 dilation sittings. Thirteen patients were lost to follow-up or did not wish to complete treatment and were excluded from the analysis. One hundred forty patients, who underwent 978 dilation sessions ( mean 6.98 per patient), completed the follow-up - mean =20.5 months (4-62). Of these, 96 were men and 44 women with a mean age of 54.1 years - median = 57yr. ( age range: 3-78 yr); 126 were white and 11, nonwhite. The etiology of the esophageal strictures of the 140 followed-up cases is shown in Table I.

Etiology versus response to therapy and number of sessions

Adequate dilation (insertion of a 42 French dilator) could be achieved in 131 of 140 (93.5%) followed up cases. The median maximum diameter of the inserted dilators were 42 French (33-51) – 14mm (11-17). The median number of sessions required for achieving adequate dilation, together with relief of dysphagia during the follow-up, among patients with peptic stenosis was 3 (range, 1-28) in comparison to 5 (range, 1-46) among those cases with postoperative strictures and also 5 (range, 1-42) among patients with caustic strictures (p=0.07, Kruskal-Wallis test).

No correlation was also found between esophageal stricture etiology and pre-treatment dysphagia, maximum diameter of the inserted bougie and post-treatment dysphagia (response to therapy), although there was a trend to a better response to therapy among patients with peptic stenosis than in patients with strictures of other causes. Mean disphagia grade after therapy was 0.83 among patients with post-operative stenosis and 0.7 among the cases with reflux esophagitis-induced stricture.

Response to therapy

Absence of dysphagia could be accomplished in 66.5% of the cases (93 of 140) at the end of follow-up. Twenty-five cases (17.8%) remained with dysphagia to some solid foods (grade 1) and 22 (15.7%) with worse grades of dysphagia – poor response. There was a strong correlation between the grade of dysphagia at presentation and the number of sessions required for adequate dilation.


Three cases of esophagus perforation in two patients with peptic stricture and in a patient with caustic stricture and 1 account of gastric perforation due to guide-wire injury in a patient with post operative stricture were observed. This represent 2.8% of the patients, but 0.40% of the sessions. If we consider the total of 153 treated cases, these figures drop to 2.6% of the cases and 0.38% of the sessions. Three of the 4 patients were managed surgically, with one death (0.7% of the cases, 0.1% of the sittings). No overt hemorrhage occurred during or after the sessions.


Although endoscopic esophageal dilation is considered the best initial therapeutic approach for benign esophageal stritures, the best technique and the best type of dilator to perform the procedure remains to be determined ( 3, 4).

Yamamoto et al. ( 7) have prospectively compared the results after Eder-Puestow or Medi-Tech balloon dilatation in 123 patients. They noted similar results with both types of dilators, whereas a previous comparison between Savary-Gilliard bougies and through-the-scope balloons also encountered no significant difference ( 8). On the other hand, Cox et al. in the final analysis of a 1-year prospective comparison of two dilation techniques found that Eder-Puestow dilators passed over an endoscopically placed guide wire produced a greater increase in esophageal lumen size at the end of the follow-up, a greater decrease in dysphagia grade at 5 months and a lower redilation rate than did nonendoscopic balloon dilation ( 9).

Every esophageal stricture seems to have its own characteristics and some individualized variation in technique may be needed for a successful therapy. Maloney dilators ( the prefered dilator type in the United States) work well for short strictures, but are difficult to employ for the initial dilation stages in more complex ones (10, 11). Eder-Puestow metal olives have been largely abandoned for Savary-Gilliard bougies in most endoscopy units around the world, although no study have prooved advantages in terms of safety and efficacy for the latter type of dilator over the former one (3, 4). In our own experience, Savary-Gilliard's bougies seem to be easier to manipulate than Eder-Puestow metal olives, although with the latter the endoscopist may "feel" better the stricture. Indeed, differences in the results of endoscopic therapy is in a large measure operator dependent, and although defined guidelines for the employment of these devices are available, interoperator variations in technique will likely result in individual biases among endoscopists and will tend to favor one type of dilator over another (3, 10, 12).

In our series, we did not employ fluoroscopic control of guide wire and dilator position and path. The use of fluoroscopy in placement and maintenance of the guide wire has been emphasized in the literature (13, 14). The American and the German Societies of Gastrointestinal Endoscopy recommend the use of fluoroscopy at least in the initial dilation sessions (2, 10). On the contrary to both societies' guidelines, Ho et al. in a prospective study evaluating 80 patients found that fluoroscopy is not necessary for Maloney dilation of benign esophageal strictures (11). Esophageal dilation using Savary-Gilliard's bougies without the aid of fluoroscopy was also reported to be safe (15). Several other recent works focused on this subject and reached similar conclusions, as the one by Marshall et al., who performed 606 wire-guided dilation sessions on 354 patients (4, 16, 17). In cases of tortuous strictures as the ones produced by lye ingestion or cancer, the passage of a very flexible guide wire of the type used to insert biliary endoprostheses can be employed as a recourse to "cannulate" these strictures with safety ( 18).

In this series, we observed one case (of a total of 4) of perforation that was due to guide wire injury to the stomach. Even with the aid of fluoroscopy, this complication would not be avoided, thus the injury occurred because the assistant did not properly hold the portion of the guide wire outside the patient – i.e., a fault in technique performance. Indeed, fluoroscopy is time-consuming, results in exposure to radiation and X ray equipment is not available in most endoscopy units because of its expense. In our opinion, fluoroscopy does not add safety to the vast majority of the procedures and even in cases of tortuous strictures, blind guide wire passage can be performed safely, when performed with care.

In our study, although not statistically significant, there was a trend towards more frequent recurrences, and consequently, dilation sessions, among patients with both post-operative and caustic strictures. Patients with peptic stenosis needed a median of 3 dilation sessions to be free of dysphagia during a mean follow-up time of 20 months, in comparison to 5 sittings among the cases with caustic or post-operative strictures. These non-statistically significant figures are probably due to a type II error.

Broor et al. found that the number of treatment sessions required to achieve adequate dilation was significantly higher among patients with corrosive (median 8, range 1 to 35) than in cases with peptic stritures (median 1, range 1 to 33). Dysphagia recurrence was also higher in the corrosive group than in the peptic stricture group (19 ). Undoubtedly, lye ingestion-induced esophageal strictures are technically more difficult to dilate and, on account of this, physicians may dilate these strictures more carefully than usual, what would result in more sessions per patient than in esophageal dilations for stricture of other etiologies.

In our series, an esophageal stenosis following surgical anastomosis was the commonest indication for endoscopic esophageal dilatation. This is a referral bias, because our hospital is a regional center for the surgical treatment of esophageal cancer. Benign anastomotic strictures after esophagectomy and cervical esophagogastrostomy or total gastrectomy and jejunoesophagostomy tend to be tighter than reflux esophagitis-induced stenoses. Although nonsignificant, thirty percent of our patients with peptic esophageal stenosis presented with mild dysphagia ( 11 of 37) in comparison to 14% ( 11 of 80) of the patients with anastomotic stricture. The vast majority of these patients were operated on for an esophageal or gastric cancer and typically dies due to their primary disease in a few months. Although recurrence should not be a problem in the long-term, due to the malignant nature of the primary disease, the tightness of their strictures usually demands more dilatation sittings at the primary setting. Our data of a median of 5 sessions needed to alleviate dysphagia is according to the wide variation found in the literature (between 2 and 9.5 sessions per patient) (20, 21, 22).

Although dysphagia caused by peptic esophageal strictures tend to be milder and easier to manage in the short-term, recurrent stricturing necessitating repeat endoscopic dilatation is a significant problem in the long-term due to the chronic nature of gastroesophageal reflux disease (23). For instance, Smith and collaborators, report that redilation was required in 30% of the patients under 20mg omeprazole a day within one year of successful treatment of the stricture (24).

Absence of dysphagia – the goal of therapy – was achieved at the end of follow-up in about two thirds of our patients, although adequate dilation was achieved in 93% of the cases; 17.8% of the cases remained with some degree of dysphagia ( grade 1), despite an adequate esophageal lumen size, indicating that a motor disorder may take part in the pathogenesis of dysphagia caused by peptic, post-operative and caustic esophageal strictures. As expected, there was a correlation between dysphagia grade at clinical presentation and the number of sessions required to achieve adequate dilation. Patients who were not able to swallow liquids in adequate amounts ( grade 4) before treatment were dilated for a median of 6 sessions in comparison to 2 and 3 sessions among patients with mild dysphagia ( grades 2 and 1, respectively).

The incidence of perforation per dilation session, in the present study (0.38%), is comparable to other large series such as the one published by a Dutch group (22) (0.4%) or by the guidelines of the German Society of Gastrointestinal Endoscopy (0.5%) (2). Other authors report more extreme results such as Marshall et al., who observed no perforation in a series of 606 wire-guided dilatations on 354 patients with hollow-core polyvinyl dilators ( 16). These authors stated that the strictures were dilated to their maximal target size in one session in 77% of the cases, what demonstrates that the majority of their patients had a mild disease. Lahoti and colleagues, dealing with corrosive strictures, which are more difficult to dilate, report 4 cases of perforation in 115 dilation sessions (3.5%) ( 25 ).

In summary, our results show that endoscopic dilatation is a safe and effective treatment for the short-term relief of dysphagia produced by benign esophageal strictures, even without the aid of fluoroscopic monitoring. In the long-term, strictures tend to recur and repeated sessions of dilation are demanded to alleviate dysphagia.

Table 1

Etiology of Benign Esophageal Strictures


Ring or web

Table 2

Dysphagia grade after endoscopic dilatation of benign esophageal strictures

Dysphagia Grade
Stricture Etiology


N %
N %
N %

27 (73)
45 (56)
6 (50)
6 (16)
17 (21)
3 (25)