- Email: [email protected]
Prospective randomized comparison of polyvinyl bougies and through-the-scope balloons for dilation of peptic strictures of the esophagus
Prospective randomized comparison of polyvinyl bougies and through-the-scope balloons for dilation of peptic strictures of the esophagus Zahid A. Saeed, MD, Carolyn B. Winchester, BSN, Pieretta S. Ferro, MD Patrice A. Michaletz, MD, Jim T. Schwartz, MD, David Y. Graham, MD Houston, Texas
We prospectively compared the efficacy of polyvinyl bougies (Savary type) passed over a guide wire and through-the-scope balloons for the dilation of peptic esophageal strictures in a randomized study. Thirty-four patients, 17 in each treatment arm, were studied. At entry, dysphagia was assessed according to a six-point scale (0, unable to swallow; 5, normal). The end-point for dilation was to size 45F or 15 mm. Discomfort during the procedure was graded on a four-point scale (0, no discomfort; 1, mild; 2, moderate; 3, severe discomfort). Follow-up visits were at I week, 1 month, 3 months, and every 3 months thereafter for 2 years. At the 1-week visit, the size of esophageal lumen was measured by 8-, 10-, and 12-mm pills. Both devices effectively relieved dysphagia. By life-table analysis, stricture recurrence during the first year of follow-up was similar in both groups, but during the second year, the risk of recurrence was significantly lower in patients whose strictures were dilated with balloons. Other advantages of balloons included the need for fewer treatment sessions to achieve the defined end-diameter for dilation (1.1 + 0.1 versus 1.7 + 0.2, p < .05), and less procedural discomfort (p < .05). The differences in luminal size after dilation, measured by the barium pill test, were not significant. Ability to pass the 12-mm pill and absence of dysphagia were correlated. Our results indicate that both devices are effective in relieving dysphagia, but balloons may have a long-term advantage. (Gastrointest Endosc 1995;41:189-95.) Historically, the mainstay of therapy for esophageal strictures has been bougienage. Since the early use of wax bougies, 1numerous improvements, including mercury-weighted rubber bougies, 2 metal olives,3 and recently, tapered polyvinyl dilators that are passed over guide wires, 4 have been introduced. All of these devices share a common mechanism of action whereby dilation is effected by the radial vector when an axially directed force encounters a stricture, analogous to the effect of a snow plow. The introduction of balloons marked the first true change in the methodology for dilating esophageal strictures. 5 With balloons, the entire dilatReceived December 10, 1994. For revision March 14, 1994. Accepted June 14, 1994. From the Gastrointestinal Endoscopy Unit, Department of Medicine, and Department of Radiology, Veterans Affairs Medical Center and Baylor College of Medicine, Houston, Texas. Reprint requests: Zahid A. Saeed, MD, Director, Gastrointestinal Endoscopy, St. Louis University Hospital (9-FDT), 3635 Vista Ave., St. Louis, MO 63110. 37/1/58897
VOLUME 41, NO. 3, 1995
ing force is delivered radially. In contrast to what occurs with bougies, if the balloon is longer than the stricture, the force is delivered simultaneously over the entire length of the stenosed segment rather than progressively from its proximal to its distal extent, a process more efficient and inherently safer from a standpoint of physics. 6 Past reports have described high degrees of success for various devices, T M and generally, one device has not been found superior to another. 15-1s Currently, perhaps the most popular method of dilation is with polyvinyl bougies passed over a guide wire (Savary). The aim of the present study was to compare prospectively through-the-scope (TTS) balloon dilators with Savary dilators in regard to efficacy and safety.
MATERIALS AND METHODS Patient selection criteria and randomization Patients evaluated for dysphagia at the Houston Veterans Affairs Medical Center between September 1987 and December 1989 were candidates for the study. All patients underwent UGI endoscopy and biopsy. Patients whose stricGASTROINTESTINAL ENDOSCOPY 1 8 9
Table 1. Scale used to assess dysphagia Numeric score
Ability to swallow
0 1
Unable to swallow Liquids with difficulty, no solids No difficulty with liquids, no solids Occasional difficulty, only with solids Rare difficulty, only with solids Normal swallowing
2 3 4 5
tures proved benign were offered participation in the study. Patients who had a history of esophageal resection (anastomotic strictures) were excluded. Patients enrolled in the study were randomly assigned to receive therapy with either Savary or T T S dilators. If the last digit of the patient's social security number was odd, therapy was with Savary, and if the last digit was even or zero, therapy was with TTS. The study was approved by the Institutional Review Boards of Baylor College of Medicine and Veterans Affairs Medical Center, Houston, Texas.
Evaluation at entry At the initial visit, each patient's ability to swallow was graded according to a six-point dysphagia scale shown in Table 1. Information was also collected regarding the presence of associated discomfort, such as heartburn, and any history of previous esophageal dilation. Endoscopic examinations were done with fiberoptic or videoendoscopes under topical oropharyngeal anesthesia and conscious sedation with intravenous diazepam (ElkinsSinn, Inc., Cherry Hill, N.J.) and butorphanol {Bristol Laboratories, Evansville, Ind.) as needed. During the initial endoscopic examination, the diameter and length of the stricture were recorded. If the stricture was not at first passable with the endoscope, its length was recorded after the dilation. T h e presence and extent of esophagitis, ulcerations, and ectopic mucosa were recorded; the presence and size of hiatus hernia were also noted. P a t i e n t discomfort during the dilation procedure was graded on the following four-point scale: 0, no discomfort; 1, mild; 2, moderate; 3, severe discomfort.
Techniques of dilation
Dilation w i t h p o l y v i n y l b o u g i e s . American Endoscopy dilators (Bard Interventional Products, Billerica, Mass.) were used for Savary dilation. If the stricture could be passed with the endoscope, a spring-tipped guide wire was placed in the antrum via the working channel of the endoscope and laid out under direct vision while the endoscope was withdrawn; otherwise, with the tip of the endoscope positioned just above the stricture, the guide wire was advanced under fluoroscopic observation through the stricture until its tip was in the antrum. The initial dilator selected was of the next size larger than the estimated stricture diameter; for example, if the stricture was thought to be 10 m m in diameter, a 33F dilator was selected. The "rule of threes ''19' 20 190
GASTROINTESTINAL ENDOSCOPY
dictated the extent of dilation during any single session (i.e., no more t h a n three dilators successively larger than the first dilator to meet resistance were passed). The defined end-point for successful dilation was the passage of a 45F dilator, and if adherence to the rule of threes precluded achieving this end-point during the initial session, dilation was completed during subsequent sessions 2 or 3 days later. T h r o u g h - t h e - s c o p e balloon dilation. Rigiflex balloons (Microvasive/Boston Scientific Corp., Watertown, Mass.) were used for T T S dilation. If the stricture could be passed with the endoscope, the balloon catheter was advanced via the working channel until the entire length of the balloon was in view. The endoscope and balloon were then withdrawn together until the balloon was in position for the dilation. If the stricture could not be passed with the endoscope, the endoscope was positioned above the stricture and the balloon catheter advanced until its tip freely cannulated the stricture. Next, it was slowly withdrawn into the stricture and inflated (with water) to its maximum pressure rating (specified by the manufacturer), typically 40 to 60 pounds per square inch (psi). The inflation was maintained for 30 seconds, released, and reinstated. After repeating this process three times, a t t e m p t s were made to move the inflated balloon through the previously strictured segment. If this was not possible, a balloon of smaller size was used before reverting to the 15-mm balloon. The size of the initial balloon selected was dictated by the estimated stricture diameter; if the stricture lumen was less than 6 mm, a 10-mm balloon was used, and if greater t h a n 6 mm, a 15-mm balloon was used. The defined end-point for dilation was the ability to move an inflated 15-mm balloon to and fro through the stricture.
Follow-up All patients were given a histamine H2-receptor antagonist (ranitidine or cimetidine), and instructions in standard antireflux measures, such as elevation of the head of the bed. Follow-up visits were at 1 week, I month, 3 months, and then every 3 months thereafter for 24 months. During these visits, the patient's ability to swallow was graded according to the six-point dysphagia scale. Additionally, patients were asked to report recurrent dysphagia if and whenever it occurred, and redilation was done with the same modality to the same end-point, as needed. At the 1-week visit, all patients underwent a barium pill test (described below) to measure objectively the effect of the dilation on the strictured esophageal lumen.
Barium pill test Barium pills 8, 10, and 12 m m in diameter were used for the study (Pharmaceutical Development Center, College of Pharmacy, Medical University of South Carolina, Charleston, S.C.). Barium pills were administered with 100 mL of water with the patient in the standing position, and their progress was followed under fluoroscopy. If the pill passed into the stomach within 60 seconds, the test for the pill of t h a t diameter was recorded as "passed." An 8-mm barium pill was given first. If the 8-mm pill passed, a 10-mm and finally a 12-mm pill were given in similar fashion. A radiograph from a patient who passed the 8 and V O L U M E 41, NO. 3, 1995
Table 2. Comparison of the clinical characteristics of patients randomized to the two treatment groups Treatment group Savary
TTS
Number of patients 17 17 Age, mean (years)* 61 + 2.1 62 + 1.7 Median 62 62 Range 43-72 43-75 Males ( %) 100 94 Previous dilation ( %)t 59 59 Dysphagia score at entry, 2.3 + 0.23 2.3+ 0.23 mean Stricture characteristics Length Mean (mm) 10.3 + 2.2 12.1 + 2.4 Median 5 5 Diameter Mean (mm) 9.9 + 0.3 8.9 + 0.7 Median 10 9 Severe esophagitis/ulcer (%) 35 59 Barrett's epithelium ( %) 12 6 Hiatus hernia (%) 82 85 Savary, Dilation with polyvinyl bougies passed over a guide wire; TTS, dilation with through-the-scope balloons. None of the differences between the two groups were statistically significant. *Mean + standard error of mean. 1-Patients with a previous history of bougienage for peptic stricture.
10 mm barium pill tests, but not the 12 mm pill test, appears in Figure 1.
Statistical analyses Wilcoxon's rank sum test and Student's t test were used for continuous variables, and Fisher's exact test (two-sided) and chi-squared analysis (with Yates' correction) were used for comparing categorical data. The Kaplan-Meier productlimit technique (life-table analysis) was used to compare the time to recurrence of dysphagia after successful dilation to 45F or 15 mm between Savary- and TTS-treated patients. "PC!info" (Retriever Data Systems, Seattle, Wash.) computer software was used to assist in the analyses. A p value of less than .05 was considered significant. Results are expressed as mean + standard error of mean.
RESULTS
Figure 1. Radiograph from a patient undergoing the barium pill test. The 8- and 10-mm barium pills passed into the stomach, where they can be seen. The 12-mm pill did not pass and is seen at the level of the stricture above the diaphragm. history of previous esophageal dilation. Strictures ranged in length from 2 to 40 m m (mean, 11.1 _+ 1.6 mm; median, 7.5 mm) and from 4 to 13 m m in diameter (mean, 9.4 _+ 0.4 ram; median, 10 mm). T h e characteristics of the patients r a n d o m i z e d to the two t r e a t m e n t groups are c o m p a r e d in Table 2. Patients in the two groups were similar u p o n e n t r y into the study.
Clinical characteristics of the patients
Achievement of the end-point
Thirty-six patients were randomized. In 2 patients the e n d - p o i n t for dilation could not be achieved. One of these patients, who was r a n d o m i z e d to receive Savary dilation and whose stricture could be dilated only to 42F, was lost to follow-up; in the other, r a n d o m i z e d to receive T T S dilation, an inflated 15-mm balloon could not be m o v e d t h r o u g h the stricture after dilation. T h e remaining 34 patients included 33 m e n and 1 w o m a n whose ages ranged from 43 to 75 years (mean, 61.5; median, 63 years). T w e n t y patients (59 % ) gave a
I t took 1.7 + 0.2 sessions (median, 2; range, 1 to 4) to achieve the e n d - p o i n t of 45F luminal diameter with Savary dilation and 1.1 _+ 0.1 sessions (median, 1; range, I to 2) to achieve 15-mm diameter with T T S dilation. This difference was statistically significant
VOLUME 41, NO. 3, 1995
(p < .05). Early subjective improvement All patients reported an i m p r o v e m e n t in their ability to swallow at the visit 1 week after dilation, GASTROINTESTINAL ENDOSCOPY 191
Table 3. Comparison of objective measurements of luminal diameter by barium pill test 1 week after dilation
1.0 L_;
O
0,8
--
Savary
L:!..........
Balloon
Pill size passed (mm)
L_. 0,6
--
L
Savary n (%) Balloon n (%)
,.°
[ O
z_
0.4
10
8
5 (31) 7 (47)
7 (44) 6 (40)
4 (25) 2 (13)
N o n e o f t h e differences were statistically significant.
--
not statistically significant. However, dysphagia was significantly less likely to be present in patients who passed the 12-mm pill as compared with those who passed only the 8-mm pill (p < .04).
0,2 - -
O
0
I
I
I
I
I
I
I
3
6 9 12 15 18 21 Time (months)
I
24
Figure 2. Life-table analysis (Kaplan-Meier product-limit technique) of time to recurrence of dysphagia. Patients treated with polyvinyl bougies (Savary) and patients treated with through-the-scope balloons (TTS) are compared. Overall, the difference was not significant (p > .05). Of the patients at risk between 12 and 24 months after dilation, significantly fewer had stricture recurrence in the TTS group (p < .05). reflected in an improvement in dysphagia scores from 2.3 + 0.16 (range, 1 to 4) before dilation to 4.5 _+ 0.13 (range, 3 to 5) after dilation. This improvement in scores was similar for patients treated with Savary dilation (2.3 +_ 0.23 to 4.5 _+ 0.16) and TTS dilation (2.3 + 0.23 to 4.4 _+ 0.21).
Objective improvement One week after dilation, 31 patients (16 Savary and 15 TTS) reported for the barium pill test. More patients whose strictures were dilated with TTS balloons passed the 12-mm barium pill (Table 3); however, the differences were not statistically significant.
Objective measurement and subjective improvement In 9 of the 12 (75 %) patients who passed the 12-mm barium pill test 1 week after dilation, the ability to swallow was restored to normal (a score of 5 according to the dysphagia scale); 2 patients had a score of 4, and i patient a score of 3. Of the 13 patients who passed the 10-ram test, 8 (62 % ) had a score of 5, 3 a score of 4, and 2 a score of 3. Of the 6 patients who passed the 8-mm tests, only 1 (17 % ) had a score of 5; 4 patients had a score of 4, and 1 a score of 3. Differences in the rates of dysphagia between patients who could pass a 12-mm or a 10-mm pill (p = .3) and between patients who could pass a 10-ram or an 8-mm pill (p = .09) were 192
12
GASTROINTESTINAL ENDOSCOPY
Time to recurrence of dysphagia Two patients in each treatment group were lost to follow-up before it could be ascertained whether dysphagia had recurred. During the 24 months of followup, 2 Savary patients and 7 TTS patients required no further dilations. No overall difference in the time to recurrence of dysphagia was seen between the two groups by life-table analysis (Fig. 2). However, if dysphagia had not recurred within 12 months after dilation, it was significantly less likely (p < .05) to recur between 12 and 24 months if the patient had undergone TTS dilation than if the patient had undergone Savary dilation (none of 7 versus 5 of 7 patients at risk in the TTS and Savary groups, respectively). Comparison was also made between patients who had a history of previous dilation and those whose strictures had not been previously dilated. Life-table analysis showed no overall difference, but during the latter 12 months of follow-up, patients with no previous history of dilation were significantly more likely (p < .05) to remain symptom-free (Fig. 3). The number of patients with a history of previous dilation in the Savary and TTS groups was similar (Table 2), and the difference between the groups for the follow-up period between 12 and 24 months did not change when the presence or absence of a history of previous dilation was taken into consideration. In patients who required repeated dilations, no difference in total number of dilation sessions was observed between the two groups (median, 3; range, I to 5 for each group).
Stricture characteristics and outcome The initial diameter of strictures that recurred was not significantly smaller than the initial diameter of strictures that did not recur (8.9 _+ 0.6 mm versus 10.2 +_ 0.4 mm, p = .08). When the entire range of initial diameters was analyzed to see whether a critical diameter would predict recurrence, no statistically significant cut-off value was found, although at a cut-off of 9 mm a strong trend (p = .07) was observed. No correlation was found between stricture length V O L U M E 41, NO. 3, 1995
(p = .6) or the presence of esophagitis (p = .7) and stricture recurrence.
A
........... P r e v i o u s k.
Discomforts and complications
During the Savary dilation to 45F diameter, 4 (24 % ) patients felt no discomfort, 6 (35 % ) felt mild, 6 (35 % ) moderate, and 1 (6 % ) severe discomfort. During T T S dilation to 15 mm, 9 (53 % ) patients felt no discomfort, 6 (35 % ) felt mild, and 2 (12 % ) moderate discomfort; no patient reported severe discomfort. According to the discomfort grading system, discomfort during TTS was significantly less (p < .05) than during Savary dilation (0.6 + 0.17 versus 1.2 + 0.22). Fluoroscopy was required for guide wire placement in 4 Savary-treated patients. In these patients, the stricture could not be passed initially with the endoscope (Olympus PQ20, diameter of 9 ram, Olympus America Inc., Lake Success, N.Y.). Fluoroscopy was not required for any TTS patient (p = .05). No procedure-related complication occurred during the course of the study. DISCUSSION
Previous studies have defined the natural history of benign strictures treated with dilation. Typically, irrespective of the type of device--bougie or balloon-used, almost all patients experience an improvement in their ability to swallow after dilation. The duration of this beneficial effect, however, is quite variable. In most patients the stricture ultimately recurs, and a need for repeated dilations has been the rule. 7-14 The present study reaffirmed these short- and long-term observations. All patients in both treatment arms experienced either complete alleviation of dysphagia or an improvement in their ability to swallow, reflected by an increase in the dysphagia score at follow-up 1 week after dilation. This initial benefit was virtually identical for both treatment arms. With time, dysphagia not unexpectedly recurred, and by 24 months of follow-up, overall only 9 (32%) patients did not require any further dilation. Because the goal of dilation, in addition to the immediate relief of dysphagia, is to provide the longestpossible symptom-free period, we used time to recurrence of dysphagia as the primary end-point for comparing the efficacy of the two devices. From past studies, one can arrive at an understanding of the significance of certain chronologic events in the natural history of strictures that bear importantly on the interpretation of results. Long-term follow-up studies have shown that the risk for requiring repeated dilation is greatest within the first year after initial dilation, 9 that dysphagia recurs in about two-thirds of the patients, and that strictures that have once required redilation are more likely to require subsequent redilations. 1° In the present study, life-table analysis confirmed these observations; most strictures recurred V O L U M E 41, NO. 3, 1995
0.8-
d ila t io n : N o
Previous dilation: Yes 1 j
: .........
"~n. o ,r-
0.6-
I---. i ...........................
I
0.4o o
I
e-
L. . . . .
0.2o o z
0
I
I
I
I
I
I
I
I
3
6
9
12
15
18
21
24
Time (months) Figure 3. Life-table analysis (Kaplan-Meier product-limit technique) of time to recurrence of dysphagia. Patients with a history of previous dilation are compared with patients without a history of previous dilation. Overall, the difference was not significant (p > .05). Of the patients at risk between 12 and 24 months after dilation, significantly fewer patients without a history of previous dilation had stricture recurrence (p < .05).
within the first year of follow-up, and the difference between previously dilated strictures and those that had been dilated for the first time became obvious during the second year (Fig. 3). For the two devices, the overall time to recurrence of dysphagia was not significantly different, but strictures dilated with TTS were significantly less likely to recur during the second year of follow-up if the first year had passed without recurrence (Fig. 2). Although the number of patients with previously dilated strictures in each treatment arm was equal (Table 2), to confirm that this difference could not be attributed to recurrences of previously dilated strictures, we performed separate life-table analyses on subgroups of patients within each treatment arm (i.e., those with previously dilated and previously undilated strictures) and found that this was not the case. However, the number of patients in each subgroup was small, and whether this difference resulted from an alpha-type error could be answered only by a considerably larger study. The end-points of dilation in the present study were the nearly equal diameters of 45F and 15 mm for Savary and TTS, respectively. For uniformity in protocol, three inflations of 30 seconds each were performed during each TTS procedure. However, in contrast to passing a bougie, inflating a balloon any given number of times does not necessarily constitute a dilation. GASTROINTESTINAL ENDOSCOPY
193
Successful TTS dilation was, therefore, recorded only if the inflated 15-mm balloon could easily be moved to and fro through the treated stricture. Compared with Savary dilation, we were able to achieve the end-point with TTS dilation in a significantly fewer number of sessions. This result was, to a great degree, influenced by initial stricture diameter and was a reflection of the guidelines used for the amount of dilation allowed during any single session. Specifically, adherence to the rule of threes resulted in a greater number of sessions required to achieve a 45F luminal diameter with Savary. Although controlled data to support the rule of threes are not available, few would argue against this time-honored precept for safe dilation. 19, 20 Guidelines for TTS, also empiric, are based on the inherent safety in the radial mechanism by which balloons effect dilation; in the present study, this proved to be an advantage in favor of TTS. In other studies, bougies and balloons of larger size have been compared. 15, 17, 18The reasons for our belief that dilating to a size larger than 45F or 15 mm is an unwarranted risk are twofold. Firstly, dilating to a size larger than 45F neither provides superior alleviation of dysphagia nor decreases recurrence rates or the need for subsequent dilations. 9 Secondly, intraesophageal pressure measurements have shown that the pressure generated during bougienage increases with increasing bougie size; in many instances the step-up in pressure is most pronounced for bougie diameters between 44F and 48F, 21 theoretically increasing the risk for perforation. It is difficult to remain totally objective in quantifying discomfort because different individuals have different thresholds for discomfort. Additionally, the amounts of sedation and analgesia, further confounded by the variations in tolerance that individuals have for a given dose of medication, have an influence. Nevertheless, the amount of sedation administered to patients in each treatment arm of the present study was similar, and given the above reservations, patients whose strictures were dilated with TTS reported significantly less discomfort during the procedure than did patients undergoing Savary dilation. The likely explanation for this observed difference in patient tolerance, in addition to the different mechanisms of act i o n - s h e a r i n g axial force for Savary versus radial for TTS6--is that the TTS procedure is entirely endoscopic, so that a patient may not even be aware of the passage of the dilator. In contrast, during Savary dilation the actual act of dilation is nonendoscopic and involves the sequential passage of several bougies, each with the potential for adding further to the level of discomfort. This observation is supported by the results of a trial in which over-the-guide wire, rather than through-the-scope, balloons were used and no difference found in patient discomfort compared with passage of bougies. 15 A further advantage of an en194
GASTROINTESTINAL ENDOSCOPY
tirely endoscopic procedure is that fluoroscopy was not required during any TTS dilation. In contrast, fluoroscopy was required during Savary procedures when the endoscope could not initially be passed beyond the stricture. Although others have advocated the blind passage of guide wires for strictures that cannot be negotiated with the endoscope, 22 we hold that the potential risks associated with blind passage are unnecessary and the practice is unsafe. 23 We used the barium pill test to determine objectively any differences between luminal diameters achieved by the two treatments. Although a few more patients undergoing TTS passed the 12-mm barium pill test, the difference between the two groups in regard to esophageal luminal diameter was not statistically significant. A significant difference was found between the rate of dysphagia for luminal diameters greater than 12 mm and the rate for diameters less than 8 mm, a finding in agreement with those of previous studies, in which correlations between luminal diameters and the degree of dysphagia caused by benign esophageal stenoses were determined by the administration of solid boluses. 24-26 The previous literature is divided between studies that show a correlation between recurrence and stricture diameter before dilation 7, 27 and those that do not. 2s In the present study, when mean initial diameter and length of a stricture were analyzed as continuous variables, mean initial diameter, but not length, was found to be correlated with recurrence. When initial diameter was analyzed categorically, an initial diameter of less than 9 mm was correlated with greater likelihood of recurrence. That these results did not achieve statistical significance is likely the result of a beta-type error. Although most benign esophageal strictures occur as a complication of gastroesophageal reflux disease, a significant number of strictures result from other causes (e.g., caustic injury or surgical anastomosis). As the natural history of peptic strictures may be different from that of other types of strictures, 29 unlike previous investigators 15, 17, is we excluded patients with nonpeptic benign strictures from this study. Because of the underlying pathophysiology of reflux disease, the presence of erosive esophagitis might have been expected to bear upon the recurrence rate. Our results show that this was not the case. Although it has been suggested that routinely prescribing an H2-receptor antagonist after dilation might eliminate any risk for recurrence caused by esophagitis, studies have failed to show a decrease in the need for repeated dilations in patients who receive such therapy. 3°, 31 In conclusion, both TTS and Savary dilation were effective in alleviating the dysphagia caused by peptic esophageal strictures. Advantages of TTS as compared with Savary included a lower recurrence rate, V O L U M E 41, NO. 3, 1995
need for fewer procedures to achieve adequate luminal diameter, and better patient tolerance. Both modalities proved safe and free of complication. Many safe and effective devices are now available to the physician to restore, at least temporarily, a patient's ability to swallow; in order to alter the natural history of strictures favorably, factors other than therapeutic devices require further investigation. REFERENCES 1. Earlam R, Cnnha-Melo JR. Benign oesophageal strictures: historical and technical aspects of dilatation. Br J Surg 1981;68: 829-36. 2. Hurst AF. Treatment of achalasia of the cardia. Q J Med 1915; 8:300-8. 3. Puestow KL. Conservative management of occlusive diseases of the esophagus. Am J Gastroenterol 1955;24:224-32. 4. Monnier P, Hsieh V, Savary M. Traitement endoscopique des stenoses oesophagiennes par les bougies de SavaryGilliard: nouveaut~s techniques. Acta Endoscopica 1985;15:11921. 5. London RL, Trotman BW, DiMarino AJ Jr, et al. Dilatation of severe esophageal strictures by an inflatable balloon catheter. Gastroenterology 1981;80:83-5. 6. McLean GK, LeVeen RF. Shear stress in the performance of esophageal dilation: comparison of balloon dilation and bougienage. Radiology 1989;172:983-6. 7. Lanza FL, Graham DY. Bougienage is effective therapy for most benign esophageal strictures. JAMA 1978;240:844-7. 8. Wesdorp ICE, Bartelsman JFWM, den Hartog Jager FCA, Huibregtse K, Tytgat GN. Results of conservative treatment of benign esophageal strictures: a follow-up study in 100 patients. Gastroenterology 1982;82:487-93. 9. Patterson DJ, Graham DY, Smith JL, et al. Natural history of benign esophageal stricture treated by dilatation. Gastroenterology 1983;85:346-50. 10. Glick ME. Clinical course of esophageal stricture managed by bougienage. Dig Dis Sci 1982;27:884-8. 11. Dumon J-F, Meric B, Sivak MV, Fleischer D. A new method of esophageal dilation using Savary-Gilliard bougies. Gastrointest Endosc 1985;31:379-82. 12. Lindor KD, Ott BJ, Hughes RW Jr. Balloon dilatation of upper digestive strictures. Gastroenterology 1985;89:545-8. 13. Graham DY, Tabibian N, Schwartz JT, Smith JL. Evaluation of the effectiveness of through-the-scope balloons as dilators of benign and malignant gastrointestinal strictures. Gastrointest Endosc 1987;33:432-6. 14. Nobrega J. Esophageal balloon dilatation: a follow-up study in 74 patients. Cardiovasc Intervent Radiol 1989;12:255-7.
VOLUME 41, NO. 3, 1995
15. Cox JGC, Winter RK, Maslin SC, et al. Balloon or bougie for dilatation of benign esophageal stricture? An interim report of a randomized controlled trial. Gut 1988;29:1741-7. 16. Michaletz PA, Saeed ZA, Winchester CB, et al. Head-to-head comparison of Savary and balloon dilatation of benign esophageal strictures: effectiveness and duration of effect [Abstract]. Gastrointest Endosc 1989;35:186. 17. Shemesh E, Czerniak A. Comparison between Savary-Gilliard and balloon dilatation of benign esophageal strictures. World J Surg 1990;14:518-22. 18. Yamamoto H, Hughes RW, Schroeder KW, Viggiano TR, DiMagno EP. Treatment of benign esophageal strictures by Eder-Peustow or balloon dilators: a comparison between randomized and prospective nonrandomized trials. Mayo Clin Proc 1992;67:228-36. 19. Boyce HW Jr. Precepts of safe esophageal dilation [Editorial]. Gastrointest Endosc 1977;23:215. 20. Tulman AB, Boyce HW Jr. Complications of esophageal dilation and guidelines for their prevention. Gastrointest Endosc 1981;27:229-34. 21. Kozarek RA, Phelps JE, Partyka EK, Sanowski RA. Intraluminal pressures generated during esophageal bougienage. Gastroenterology 1981;81:833-7. 22. Kadakia SC, Cohan CF, Starnes EC. Esophageal dilation with polyvinyl bougies using a guide wire with markings without the aid of fluoroscopy. Gastrointest Endosc 1991;37:183-6. 23. Graham DY, Saeed ZA. Guide wire-assisted esophageal dilation without fluoroscopy: truly blind dilation is almost always a mistake. Gastrointest Endosc 1991;37:650-1. 24. Schatzki R. The lower esophageal ring: long-term follow-up of symptomatic and asymptomatic rings. Am J Radiol 1963;90: 805-10 25. Goldschmid S, Boyce HW Jr, Brown JI, Brady PG, Nord HJ, Lyman GH. A new objective measurement of esophageal lumen patency. Am J Gastroenterol 1989;84:1255-8. 26. Ott DJ, Kelley TF, Chen MYM, Gelfand DW, Wu WC. Use of a marshmallow bolus for evaluating lower esophageal mucosal rings. Am J Gastroenterol 1991;86:817-20. 27. Fellows IW, RainaS, Holmes GKT. Celestindilatationofbenign esophageal strictures: a review of 100 patients. Am J Gastroenterol 1986;81:1052-4. 28. Ogilvie PL, Ferguson R, Atkinson M. Outlook with conservative treatment of peptic oesophageal stricture. Gut 1980;240:844-7. 29. Webb WA, McDaniel L, Jones L. Endoscopic evaluation of dysphagia in two hundred and ninety-three patients with benign disease. Surg Gynecol Obstet 1984;158:152-6. 30. Ferguson R, Dronfield MW, Atkinson M. Cimetidine in the treatment of reflux oesophagitis with peptic strictures. BMJ 1979;2:472-4. 31. 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;27:594-8.
GASTROINTESTINAL ENDOSCOPY 195
We prospectively compared the efficacy of polyvinyl bougies (Savary type) passed over a guide wire and through-the-scope balloons for the dilation of peptic esophageal strictures in a randomized study. Thirty-four patients, 17 in each treatment arm, were studied. At entry, dysphagia was assessed according to a six-point scale (0, unable to swallow; 5, normal). The end-point for dilation was to size 45F or 15 mm. Discomfort during the procedure was graded on a four-point scale (0, no discomfort; 1, mild; 2, moderate; 3, severe discomfort). Follow-up visits were at I week, 1 month, 3 months, and every 3 months thereafter for 2 years. At the 1-week visit, the size of esophageal lumen was measured by 8-, 10-, and 12-mm pills. Both devices effectively relieved dysphagia. By life-table analysis, stricture recurrence during the first year of follow-up was similar in both groups, but during the second year, the risk of recurrence was significantly lower in patients whose strictures were dilated with balloons. Other advantages of balloons included the need for fewer treatment sessions to achieve the defined end-diameter for dilation (1.1 + 0.1 versus 1.7 + 0.2, p < .05), and less procedural discomfort (p < .05). The differences in luminal size after dilation, measured by the barium pill test, were not significant. Ability to pass the 12-mm pill and absence of dysphagia were correlated. Our results indicate that both devices are effective in relieving dysphagia, but balloons may have a long-term advantage. (Gastrointest Endosc 1995;41:189-95.) Historically, the mainstay of therapy for esophageal strictures has been bougienage. Since the early use of wax bougies, 1numerous improvements, including mercury-weighted rubber bougies, 2 metal olives,3 and recently, tapered polyvinyl dilators that are passed over guide wires, 4 have been introduced. All of these devices share a common mechanism of action whereby dilation is effected by the radial vector when an axially directed force encounters a stricture, analogous to the effect of a snow plow. The introduction of balloons marked the first true change in the methodology for dilating esophageal strictures. 5 With balloons, the entire dilatReceived December 10, 1994. For revision March 14, 1994. Accepted June 14, 1994. From the Gastrointestinal Endoscopy Unit, Department of Medicine, and Department of Radiology, Veterans Affairs Medical Center and Baylor College of Medicine, Houston, Texas. Reprint requests: Zahid A. Saeed, MD, Director, Gastrointestinal Endoscopy, St. Louis University Hospital (9-FDT), 3635 Vista Ave., St. Louis, MO 63110. 37/1/58897
VOLUME 41, NO. 3, 1995
ing force is delivered radially. In contrast to what occurs with bougies, if the balloon is longer than the stricture, the force is delivered simultaneously over the entire length of the stenosed segment rather than progressively from its proximal to its distal extent, a process more efficient and inherently safer from a standpoint of physics. 6 Past reports have described high degrees of success for various devices, T M and generally, one device has not been found superior to another. 15-1s Currently, perhaps the most popular method of dilation is with polyvinyl bougies passed over a guide wire (Savary). The aim of the present study was to compare prospectively through-the-scope (TTS) balloon dilators with Savary dilators in regard to efficacy and safety.
MATERIALS AND METHODS Patient selection criteria and randomization Patients evaluated for dysphagia at the Houston Veterans Affairs Medical Center between September 1987 and December 1989 were candidates for the study. All patients underwent UGI endoscopy and biopsy. Patients whose stricGASTROINTESTINAL ENDOSCOPY 1 8 9
Table 1. Scale used to assess dysphagia Numeric score
Ability to swallow
0 1
Unable to swallow Liquids with difficulty, no solids No difficulty with liquids, no solids Occasional difficulty, only with solids Rare difficulty, only with solids Normal swallowing
2 3 4 5
tures proved benign were offered participation in the study. Patients who had a history of esophageal resection (anastomotic strictures) were excluded. Patients enrolled in the study were randomly assigned to receive therapy with either Savary or T T S dilators. If the last digit of the patient's social security number was odd, therapy was with Savary, and if the last digit was even or zero, therapy was with TTS. The study was approved by the Institutional Review Boards of Baylor College of Medicine and Veterans Affairs Medical Center, Houston, Texas.
Evaluation at entry At the initial visit, each patient's ability to swallow was graded according to a six-point dysphagia scale shown in Table 1. Information was also collected regarding the presence of associated discomfort, such as heartburn, and any history of previous esophageal dilation. Endoscopic examinations were done with fiberoptic or videoendoscopes under topical oropharyngeal anesthesia and conscious sedation with intravenous diazepam (ElkinsSinn, Inc., Cherry Hill, N.J.) and butorphanol {Bristol Laboratories, Evansville, Ind.) as needed. During the initial endoscopic examination, the diameter and length of the stricture were recorded. If the stricture was not at first passable with the endoscope, its length was recorded after the dilation. T h e presence and extent of esophagitis, ulcerations, and ectopic mucosa were recorded; the presence and size of hiatus hernia were also noted. P a t i e n t discomfort during the dilation procedure was graded on the following four-point scale: 0, no discomfort; 1, mild; 2, moderate; 3, severe discomfort.
Techniques of dilation
Dilation w i t h p o l y v i n y l b o u g i e s . American Endoscopy dilators (Bard Interventional Products, Billerica, Mass.) were used for Savary dilation. If the stricture could be passed with the endoscope, a spring-tipped guide wire was placed in the antrum via the working channel of the endoscope and laid out under direct vision while the endoscope was withdrawn; otherwise, with the tip of the endoscope positioned just above the stricture, the guide wire was advanced under fluoroscopic observation through the stricture until its tip was in the antrum. The initial dilator selected was of the next size larger than the estimated stricture diameter; for example, if the stricture was thought to be 10 m m in diameter, a 33F dilator was selected. The "rule of threes ''19' 20 190
GASTROINTESTINAL ENDOSCOPY
dictated the extent of dilation during any single session (i.e., no more t h a n three dilators successively larger than the first dilator to meet resistance were passed). The defined end-point for successful dilation was the passage of a 45F dilator, and if adherence to the rule of threes precluded achieving this end-point during the initial session, dilation was completed during subsequent sessions 2 or 3 days later. T h r o u g h - t h e - s c o p e balloon dilation. Rigiflex balloons (Microvasive/Boston Scientific Corp., Watertown, Mass.) were used for T T S dilation. If the stricture could be passed with the endoscope, the balloon catheter was advanced via the working channel until the entire length of the balloon was in view. The endoscope and balloon were then withdrawn together until the balloon was in position for the dilation. If the stricture could not be passed with the endoscope, the endoscope was positioned above the stricture and the balloon catheter advanced until its tip freely cannulated the stricture. Next, it was slowly withdrawn into the stricture and inflated (with water) to its maximum pressure rating (specified by the manufacturer), typically 40 to 60 pounds per square inch (psi). The inflation was maintained for 30 seconds, released, and reinstated. After repeating this process three times, a t t e m p t s were made to move the inflated balloon through the previously strictured segment. If this was not possible, a balloon of smaller size was used before reverting to the 15-mm balloon. The size of the initial balloon selected was dictated by the estimated stricture diameter; if the stricture lumen was less than 6 mm, a 10-mm balloon was used, and if greater t h a n 6 mm, a 15-mm balloon was used. The defined end-point for dilation was the ability to move an inflated 15-mm balloon to and fro through the stricture.
Follow-up All patients were given a histamine H2-receptor antagonist (ranitidine or cimetidine), and instructions in standard antireflux measures, such as elevation of the head of the bed. Follow-up visits were at 1 week, I month, 3 months, and then every 3 months thereafter for 24 months. During these visits, the patient's ability to swallow was graded according to the six-point dysphagia scale. Additionally, patients were asked to report recurrent dysphagia if and whenever it occurred, and redilation was done with the same modality to the same end-point, as needed. At the 1-week visit, all patients underwent a barium pill test (described below) to measure objectively the effect of the dilation on the strictured esophageal lumen.
Barium pill test Barium pills 8, 10, and 12 m m in diameter were used for the study (Pharmaceutical Development Center, College of Pharmacy, Medical University of South Carolina, Charleston, S.C.). Barium pills were administered with 100 mL of water with the patient in the standing position, and their progress was followed under fluoroscopy. If the pill passed into the stomach within 60 seconds, the test for the pill of t h a t diameter was recorded as "passed." An 8-mm barium pill was given first. If the 8-mm pill passed, a 10-mm and finally a 12-mm pill were given in similar fashion. A radiograph from a patient who passed the 8 and V O L U M E 41, NO. 3, 1995
Table 2. Comparison of the clinical characteristics of patients randomized to the two treatment groups Treatment group Savary
TTS
Number of patients 17 17 Age, mean (years)* 61 + 2.1 62 + 1.7 Median 62 62 Range 43-72 43-75 Males ( %) 100 94 Previous dilation ( %)t 59 59 Dysphagia score at entry, 2.3 + 0.23 2.3+ 0.23 mean Stricture characteristics Length Mean (mm) 10.3 + 2.2 12.1 + 2.4 Median 5 5 Diameter Mean (mm) 9.9 + 0.3 8.9 + 0.7 Median 10 9 Severe esophagitis/ulcer (%) 35 59 Barrett's epithelium ( %) 12 6 Hiatus hernia (%) 82 85 Savary, Dilation with polyvinyl bougies passed over a guide wire; TTS, dilation with through-the-scope balloons. None of the differences between the two groups were statistically significant. *Mean + standard error of mean. 1-Patients with a previous history of bougienage for peptic stricture.
10 mm barium pill tests, but not the 12 mm pill test, appears in Figure 1.
Statistical analyses Wilcoxon's rank sum test and Student's t test were used for continuous variables, and Fisher's exact test (two-sided) and chi-squared analysis (with Yates' correction) were used for comparing categorical data. The Kaplan-Meier productlimit technique (life-table analysis) was used to compare the time to recurrence of dysphagia after successful dilation to 45F or 15 mm between Savary- and TTS-treated patients. "PC!info" (Retriever Data Systems, Seattle, Wash.) computer software was used to assist in the analyses. A p value of less than .05 was considered significant. Results are expressed as mean + standard error of mean.
RESULTS
Figure 1. Radiograph from a patient undergoing the barium pill test. The 8- and 10-mm barium pills passed into the stomach, where they can be seen. The 12-mm pill did not pass and is seen at the level of the stricture above the diaphragm. history of previous esophageal dilation. Strictures ranged in length from 2 to 40 m m (mean, 11.1 _+ 1.6 mm; median, 7.5 mm) and from 4 to 13 m m in diameter (mean, 9.4 _+ 0.4 ram; median, 10 mm). T h e characteristics of the patients r a n d o m i z e d to the two t r e a t m e n t groups are c o m p a r e d in Table 2. Patients in the two groups were similar u p o n e n t r y into the study.
Clinical characteristics of the patients
Achievement of the end-point
Thirty-six patients were randomized. In 2 patients the e n d - p o i n t for dilation could not be achieved. One of these patients, who was r a n d o m i z e d to receive Savary dilation and whose stricture could be dilated only to 42F, was lost to follow-up; in the other, r a n d o m i z e d to receive T T S dilation, an inflated 15-mm balloon could not be m o v e d t h r o u g h the stricture after dilation. T h e remaining 34 patients included 33 m e n and 1 w o m a n whose ages ranged from 43 to 75 years (mean, 61.5; median, 63 years). T w e n t y patients (59 % ) gave a
I t took 1.7 + 0.2 sessions (median, 2; range, 1 to 4) to achieve the e n d - p o i n t of 45F luminal diameter with Savary dilation and 1.1 _+ 0.1 sessions (median, 1; range, I to 2) to achieve 15-mm diameter with T T S dilation. This difference was statistically significant
VOLUME 41, NO. 3, 1995
(p < .05). Early subjective improvement All patients reported an i m p r o v e m e n t in their ability to swallow at the visit 1 week after dilation, GASTROINTESTINAL ENDOSCOPY 191
Table 3. Comparison of objective measurements of luminal diameter by barium pill test 1 week after dilation
1.0 L_;
O
0,8
--
Savary
L:!..........
Balloon
Pill size passed (mm)
L_. 0,6
--
L
Savary n (%) Balloon n (%)
,.°
[ O
z_
0.4
10
8
5 (31) 7 (47)
7 (44) 6 (40)
4 (25) 2 (13)
N o n e o f t h e differences were statistically significant.
--
not statistically significant. However, dysphagia was significantly less likely to be present in patients who passed the 12-mm pill as compared with those who passed only the 8-mm pill (p < .04).
0,2 - -
O
0
I
I
I
I
I
I
I
3
6 9 12 15 18 21 Time (months)
I
24
Figure 2. Life-table analysis (Kaplan-Meier product-limit technique) of time to recurrence of dysphagia. Patients treated with polyvinyl bougies (Savary) and patients treated with through-the-scope balloons (TTS) are compared. Overall, the difference was not significant (p > .05). Of the patients at risk between 12 and 24 months after dilation, significantly fewer had stricture recurrence in the TTS group (p < .05). reflected in an improvement in dysphagia scores from 2.3 + 0.16 (range, 1 to 4) before dilation to 4.5 _+ 0.13 (range, 3 to 5) after dilation. This improvement in scores was similar for patients treated with Savary dilation (2.3 +_ 0.23 to 4.5 _+ 0.16) and TTS dilation (2.3 + 0.23 to 4.4 _+ 0.21).
Objective improvement One week after dilation, 31 patients (16 Savary and 15 TTS) reported for the barium pill test. More patients whose strictures were dilated with TTS balloons passed the 12-mm barium pill (Table 3); however, the differences were not statistically significant.
Objective measurement and subjective improvement In 9 of the 12 (75 %) patients who passed the 12-mm barium pill test 1 week after dilation, the ability to swallow was restored to normal (a score of 5 according to the dysphagia scale); 2 patients had a score of 4, and i patient a score of 3. Of the 13 patients who passed the 10-ram test, 8 (62 % ) had a score of 5, 3 a score of 4, and 2 a score of 3. Of the 6 patients who passed the 8-mm tests, only 1 (17 % ) had a score of 5; 4 patients had a score of 4, and 1 a score of 3. Differences in the rates of dysphagia between patients who could pass a 12-mm or a 10-mm pill (p = .3) and between patients who could pass a 10-ram or an 8-mm pill (p = .09) were 192
12
GASTROINTESTINAL ENDOSCOPY
Time to recurrence of dysphagia Two patients in each treatment group were lost to follow-up before it could be ascertained whether dysphagia had recurred. During the 24 months of followup, 2 Savary patients and 7 TTS patients required no further dilations. No overall difference in the time to recurrence of dysphagia was seen between the two groups by life-table analysis (Fig. 2). However, if dysphagia had not recurred within 12 months after dilation, it was significantly less likely (p < .05) to recur between 12 and 24 months if the patient had undergone TTS dilation than if the patient had undergone Savary dilation (none of 7 versus 5 of 7 patients at risk in the TTS and Savary groups, respectively). Comparison was also made between patients who had a history of previous dilation and those whose strictures had not been previously dilated. Life-table analysis showed no overall difference, but during the latter 12 months of follow-up, patients with no previous history of dilation were significantly more likely (p < .05) to remain symptom-free (Fig. 3). The number of patients with a history of previous dilation in the Savary and TTS groups was similar (Table 2), and the difference between the groups for the follow-up period between 12 and 24 months did not change when the presence or absence of a history of previous dilation was taken into consideration. In patients who required repeated dilations, no difference in total number of dilation sessions was observed between the two groups (median, 3; range, I to 5 for each group).
Stricture characteristics and outcome The initial diameter of strictures that recurred was not significantly smaller than the initial diameter of strictures that did not recur (8.9 _+ 0.6 mm versus 10.2 +_ 0.4 mm, p = .08). When the entire range of initial diameters was analyzed to see whether a critical diameter would predict recurrence, no statistically significant cut-off value was found, although at a cut-off of 9 mm a strong trend (p = .07) was observed. No correlation was found between stricture length V O L U M E 41, NO. 3, 1995
(p = .6) or the presence of esophagitis (p = .7) and stricture recurrence.
A
........... P r e v i o u s k.
Discomforts and complications
During the Savary dilation to 45F diameter, 4 (24 % ) patients felt no discomfort, 6 (35 % ) felt mild, 6 (35 % ) moderate, and 1 (6 % ) severe discomfort. During T T S dilation to 15 mm, 9 (53 % ) patients felt no discomfort, 6 (35 % ) felt mild, and 2 (12 % ) moderate discomfort; no patient reported severe discomfort. According to the discomfort grading system, discomfort during TTS was significantly less (p < .05) than during Savary dilation (0.6 + 0.17 versus 1.2 + 0.22). Fluoroscopy was required for guide wire placement in 4 Savary-treated patients. In these patients, the stricture could not be passed initially with the endoscope (Olympus PQ20, diameter of 9 ram, Olympus America Inc., Lake Success, N.Y.). Fluoroscopy was not required for any TTS patient (p = .05). No procedure-related complication occurred during the course of the study. DISCUSSION
Previous studies have defined the natural history of benign strictures treated with dilation. Typically, irrespective of the type of device--bougie or balloon-used, almost all patients experience an improvement in their ability to swallow after dilation. The duration of this beneficial effect, however, is quite variable. In most patients the stricture ultimately recurs, and a need for repeated dilations has been the rule. 7-14 The present study reaffirmed these short- and long-term observations. All patients in both treatment arms experienced either complete alleviation of dysphagia or an improvement in their ability to swallow, reflected by an increase in the dysphagia score at follow-up 1 week after dilation. This initial benefit was virtually identical for both treatment arms. With time, dysphagia not unexpectedly recurred, and by 24 months of follow-up, overall only 9 (32%) patients did not require any further dilation. Because the goal of dilation, in addition to the immediate relief of dysphagia, is to provide the longestpossible symptom-free period, we used time to recurrence of dysphagia as the primary end-point for comparing the efficacy of the two devices. From past studies, one can arrive at an understanding of the significance of certain chronologic events in the natural history of strictures that bear importantly on the interpretation of results. Long-term follow-up studies have shown that the risk for requiring repeated dilation is greatest within the first year after initial dilation, 9 that dysphagia recurs in about two-thirds of the patients, and that strictures that have once required redilation are more likely to require subsequent redilations. 1° In the present study, life-table analysis confirmed these observations; most strictures recurred V O L U M E 41, NO. 3, 1995
0.8-
d ila t io n : N o
Previous dilation: Yes 1 j
: .........
"~n. o ,r-
0.6-
I---. i ...........................
I
0.4o o
I
e-
L. . . . .
0.2o o z
0
I
I
I
I
I
I
I
I
3
6
9
12
15
18
21
24
Time (months) Figure 3. Life-table analysis (Kaplan-Meier product-limit technique) of time to recurrence of dysphagia. Patients with a history of previous dilation are compared with patients without a history of previous dilation. Overall, the difference was not significant (p > .05). Of the patients at risk between 12 and 24 months after dilation, significantly fewer patients without a history of previous dilation had stricture recurrence (p < .05).
within the first year of follow-up, and the difference between previously dilated strictures and those that had been dilated for the first time became obvious during the second year (Fig. 3). For the two devices, the overall time to recurrence of dysphagia was not significantly different, but strictures dilated with TTS were significantly less likely to recur during the second year of follow-up if the first year had passed without recurrence (Fig. 2). Although the number of patients with previously dilated strictures in each treatment arm was equal (Table 2), to confirm that this difference could not be attributed to recurrences of previously dilated strictures, we performed separate life-table analyses on subgroups of patients within each treatment arm (i.e., those with previously dilated and previously undilated strictures) and found that this was not the case. However, the number of patients in each subgroup was small, and whether this difference resulted from an alpha-type error could be answered only by a considerably larger study. The end-points of dilation in the present study were the nearly equal diameters of 45F and 15 mm for Savary and TTS, respectively. For uniformity in protocol, three inflations of 30 seconds each were performed during each TTS procedure. However, in contrast to passing a bougie, inflating a balloon any given number of times does not necessarily constitute a dilation. GASTROINTESTINAL ENDOSCOPY
193
Successful TTS dilation was, therefore, recorded only if the inflated 15-mm balloon could easily be moved to and fro through the treated stricture. Compared with Savary dilation, we were able to achieve the end-point with TTS dilation in a significantly fewer number of sessions. This result was, to a great degree, influenced by initial stricture diameter and was a reflection of the guidelines used for the amount of dilation allowed during any single session. Specifically, adherence to the rule of threes resulted in a greater number of sessions required to achieve a 45F luminal diameter with Savary. Although controlled data to support the rule of threes are not available, few would argue against this time-honored precept for safe dilation. 19, 20 Guidelines for TTS, also empiric, are based on the inherent safety in the radial mechanism by which balloons effect dilation; in the present study, this proved to be an advantage in favor of TTS. In other studies, bougies and balloons of larger size have been compared. 15, 17, 18The reasons for our belief that dilating to a size larger than 45F or 15 mm is an unwarranted risk are twofold. Firstly, dilating to a size larger than 45F neither provides superior alleviation of dysphagia nor decreases recurrence rates or the need for subsequent dilations. 9 Secondly, intraesophageal pressure measurements have shown that the pressure generated during bougienage increases with increasing bougie size; in many instances the step-up in pressure is most pronounced for bougie diameters between 44F and 48F, 21 theoretically increasing the risk for perforation. It is difficult to remain totally objective in quantifying discomfort because different individuals have different thresholds for discomfort. Additionally, the amounts of sedation and analgesia, further confounded by the variations in tolerance that individuals have for a given dose of medication, have an influence. Nevertheless, the amount of sedation administered to patients in each treatment arm of the present study was similar, and given the above reservations, patients whose strictures were dilated with TTS reported significantly less discomfort during the procedure than did patients undergoing Savary dilation. The likely explanation for this observed difference in patient tolerance, in addition to the different mechanisms of act i o n - s h e a r i n g axial force for Savary versus radial for TTS6--is that the TTS procedure is entirely endoscopic, so that a patient may not even be aware of the passage of the dilator. In contrast, during Savary dilation the actual act of dilation is nonendoscopic and involves the sequential passage of several bougies, each with the potential for adding further to the level of discomfort. This observation is supported by the results of a trial in which over-the-guide wire, rather than through-the-scope, balloons were used and no difference found in patient discomfort compared with passage of bougies. 15 A further advantage of an en194
GASTROINTESTINAL ENDOSCOPY
tirely endoscopic procedure is that fluoroscopy was not required during any TTS dilation. In contrast, fluoroscopy was required during Savary procedures when the endoscope could not initially be passed beyond the stricture. Although others have advocated the blind passage of guide wires for strictures that cannot be negotiated with the endoscope, 22 we hold that the potential risks associated with blind passage are unnecessary and the practice is unsafe. 23 We used the barium pill test to determine objectively any differences between luminal diameters achieved by the two treatments. Although a few more patients undergoing TTS passed the 12-mm barium pill test, the difference between the two groups in regard to esophageal luminal diameter was not statistically significant. A significant difference was found between the rate of dysphagia for luminal diameters greater than 12 mm and the rate for diameters less than 8 mm, a finding in agreement with those of previous studies, in which correlations between luminal diameters and the degree of dysphagia caused by benign esophageal stenoses were determined by the administration of solid boluses. 24-26 The previous literature is divided between studies that show a correlation between recurrence and stricture diameter before dilation 7, 27 and those that do not. 2s In the present study, when mean initial diameter and length of a stricture were analyzed as continuous variables, mean initial diameter, but not length, was found to be correlated with recurrence. When initial diameter was analyzed categorically, an initial diameter of less than 9 mm was correlated with greater likelihood of recurrence. That these results did not achieve statistical significance is likely the result of a beta-type error. Although most benign esophageal strictures occur as a complication of gastroesophageal reflux disease, a significant number of strictures result from other causes (e.g., caustic injury or surgical anastomosis). As the natural history of peptic strictures may be different from that of other types of strictures, 29 unlike previous investigators 15, 17, is we excluded patients with nonpeptic benign strictures from this study. Because of the underlying pathophysiology of reflux disease, the presence of erosive esophagitis might have been expected to bear upon the recurrence rate. Our results show that this was not the case. Although it has been suggested that routinely prescribing an H2-receptor antagonist after dilation might eliminate any risk for recurrence caused by esophagitis, studies have failed to show a decrease in the need for repeated dilations in patients who receive such therapy. 3°, 31 In conclusion, both TTS and Savary dilation were effective in alleviating the dysphagia caused by peptic esophageal strictures. Advantages of TTS as compared with Savary included a lower recurrence rate, V O L U M E 41, NO. 3, 1995
need for fewer procedures to achieve adequate luminal diameter, and better patient tolerance. Both modalities proved safe and free of complication. Many safe and effective devices are now available to the physician to restore, at least temporarily, a patient's ability to swallow; in order to alter the natural history of strictures favorably, factors other than therapeutic devices require further investigation. REFERENCES 1. Earlam R, Cnnha-Melo JR. Benign oesophageal strictures: historical and technical aspects of dilatation. Br J Surg 1981;68: 829-36. 2. Hurst AF. Treatment of achalasia of the cardia. Q J Med 1915; 8:300-8. 3. Puestow KL. Conservative management of occlusive diseases of the esophagus. Am J Gastroenterol 1955;24:224-32. 4. Monnier P, Hsieh V, Savary M. Traitement endoscopique des stenoses oesophagiennes par les bougies de SavaryGilliard: nouveaut~s techniques. Acta Endoscopica 1985;15:11921. 5. London RL, Trotman BW, DiMarino AJ Jr, et al. Dilatation of severe esophageal strictures by an inflatable balloon catheter. Gastroenterology 1981;80:83-5. 6. McLean GK, LeVeen RF. Shear stress in the performance of esophageal dilation: comparison of balloon dilation and bougienage. Radiology 1989;172:983-6. 7. Lanza FL, Graham DY. Bougienage is effective therapy for most benign esophageal strictures. JAMA 1978;240:844-7. 8. Wesdorp ICE, Bartelsman JFWM, den Hartog Jager FCA, Huibregtse K, Tytgat GN. Results of conservative treatment of benign esophageal strictures: a follow-up study in 100 patients. Gastroenterology 1982;82:487-93. 9. Patterson DJ, Graham DY, Smith JL, et al. Natural history of benign esophageal stricture treated by dilatation. Gastroenterology 1983;85:346-50. 10. Glick ME. Clinical course of esophageal stricture managed by bougienage. Dig Dis Sci 1982;27:884-8. 11. Dumon J-F, Meric B, Sivak MV, Fleischer D. A new method of esophageal dilation using Savary-Gilliard bougies. Gastrointest Endosc 1985;31:379-82. 12. Lindor KD, Ott BJ, Hughes RW Jr. Balloon dilatation of upper digestive strictures. Gastroenterology 1985;89:545-8. 13. Graham DY, Tabibian N, Schwartz JT, Smith JL. Evaluation of the effectiveness of through-the-scope balloons as dilators of benign and malignant gastrointestinal strictures. Gastrointest Endosc 1987;33:432-6. 14. Nobrega J. Esophageal balloon dilatation: a follow-up study in 74 patients. Cardiovasc Intervent Radiol 1989;12:255-7.
VOLUME 41, NO. 3, 1995
15. Cox JGC, Winter RK, Maslin SC, et al. Balloon or bougie for dilatation of benign esophageal stricture? An interim report of a randomized controlled trial. Gut 1988;29:1741-7. 16. Michaletz PA, Saeed ZA, Winchester CB, et al. Head-to-head comparison of Savary and balloon dilatation of benign esophageal strictures: effectiveness and duration of effect [Abstract]. Gastrointest Endosc 1989;35:186. 17. Shemesh E, Czerniak A. Comparison between Savary-Gilliard and balloon dilatation of benign esophageal strictures. World J Surg 1990;14:518-22. 18. Yamamoto H, Hughes RW, Schroeder KW, Viggiano TR, DiMagno EP. Treatment of benign esophageal strictures by Eder-Peustow or balloon dilators: a comparison between randomized and prospective nonrandomized trials. Mayo Clin Proc 1992;67:228-36. 19. Boyce HW Jr. Precepts of safe esophageal dilation [Editorial]. Gastrointest Endosc 1977;23:215. 20. Tulman AB, Boyce HW Jr. Complications of esophageal dilation and guidelines for their prevention. Gastrointest Endosc 1981;27:229-34. 21. Kozarek RA, Phelps JE, Partyka EK, Sanowski RA. Intraluminal pressures generated during esophageal bougienage. Gastroenterology 1981;81:833-7. 22. Kadakia SC, Cohan CF, Starnes EC. Esophageal dilation with polyvinyl bougies using a guide wire with markings without the aid of fluoroscopy. Gastrointest Endosc 1991;37:183-6. 23. Graham DY, Saeed ZA. Guide wire-assisted esophageal dilation without fluoroscopy: truly blind dilation is almost always a mistake. Gastrointest Endosc 1991;37:650-1. 24. Schatzki R. The lower esophageal ring: long-term follow-up of symptomatic and asymptomatic rings. Am J Radiol 1963;90: 805-10 25. Goldschmid S, Boyce HW Jr, Brown JI, Brady PG, Nord HJ, Lyman GH. A new objective measurement of esophageal lumen patency. Am J Gastroenterol 1989;84:1255-8. 26. Ott DJ, Kelley TF, Chen MYM, Gelfand DW, Wu WC. Use of a marshmallow bolus for evaluating lower esophageal mucosal rings. Am J Gastroenterol 1991;86:817-20. 27. Fellows IW, RainaS, Holmes GKT. Celestindilatationofbenign esophageal strictures: a review of 100 patients. Am J Gastroenterol 1986;81:1052-4. 28. Ogilvie PL, Ferguson R, Atkinson M. Outlook with conservative treatment of peptic oesophageal stricture. Gut 1980;240:844-7. 29. Webb WA, McDaniel L, Jones L. Endoscopic evaluation of dysphagia in two hundred and ninety-three patients with benign disease. Surg Gynecol Obstet 1984;158:152-6. 30. Ferguson R, Dronfield MW, Atkinson M. Cimetidine in the treatment of reflux oesophagitis with peptic strictures. BMJ 1979;2:472-4. 31. 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;27:594-8.
GASTROINTESTINAL ENDOSCOPY 195