| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Radiology, Hospital of the University of Pennsylvania, 3400 Spruce St., Philadelphia, PA 19104.
Received December 16, 2002;
accepted after revision February 5, 2003.
Address correspondence to M. S. Levine
(levine{at}oasis.rad.upenn.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS.A search of our radiology database from 1998 to 2001 revealed 46 eligible radiographic studies performed using a water-soluble contrast agent alone or a water-soluble contrast agent followed by barium that showed leaks in patients who had undergone esophagogastrectomy, total gastrectomy, or total laryngectomy. The images were reviewed to determine the morphology of the leaks (i.e., blind-ending tracks, sealed-off collections, or free extravasation of contrast material). Medical records were also reviewed to determine whether detection of the leaks seen on the radiographic studies affected patient management.
RESULTS. Of the 46 leaks seen on radiographic studies, 23 (50%) were detected with a water-soluble contrast agent and 23 (50%) were detected only with high-density barium. Of the 23 leaks visualized with water-soluble contrast media, six (26%) were characterized by blind-ending tracks, 14 (61%) by sealed-off collections, and three (13%) by free extravasation of contrast material into the mediastinum or neck. Of the 23 leaks visualized only with high-density barium, 19 (83%) were characterized by blind-ending tracks and four (17%) by sealed-off collections. Thus, leaks detected only on images obtained with high-density barium were significantly more likely to be characterized by blind-ending tracks than those detected on images obtained with a water-soluble contrast agent (p = 0.0007). Of the 33 patients with clinical follow-up, the findings seen on these imaging studies affected management in 12 (86%) of 14 patients with leaks depicted by water-soluble contrast media and in 10 (53%) of 19 with leaks depicted only by high-density barium.
CONCLUSION. Our findings support the use of high-density barium as part of the routine postoperative radiographic examination when no leaks are detected on images obtained with a water-soluble contrast agent.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Two more recent investigations that compared water-soluble contrast agents with barium for the detection of esophageal leaks have corroborated these earlier observations. In a study by Tanomkiat and Galassi [8], 14 (38%) of 37 leaks missed on images obtained with a water-soluble contrast agent were visualized on images obtained with a 60% weight/volume (w/v) barium suspension. In another study by Buecker et al. [9], four (22%) of 18 leaks missed with a water-soluble contrast agent were visualized with a 100% w/v barium suspension. Most authors therefore believe that the radiographic examination should be repeated immediately with barium if no leak is seen with a water-soluble contrast agent in patients with a possible esophageal perforation [1–4, 8, 9].
Because high-density barium suspensions are more radiopaque than low-density barium suspensions and more adherent to sites of extraluminal leakage, we have believed for some time that high-density barium is even more effective than low-density barium for showing perforations in patients with pharyngeal or esophageal leaks. As a result, we routinely use a 250% w/v barium suspension (the same barium used for routine double-contrast upper gastrointestinal tract examinations) for radiographic evaluation of possible pharyngeal or esophageal perforations if no leak is shown on images obtained with water-soluble contrast medium. The purpose of this investigation was to determine the usefulness of a 250% w/v barium suspension compared with a water-soluble contrast agent for the detection of leaks in patients who have undergone esophagogastrectomy, total gastrectomy, or total laryngectomy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All 46 patients initially were given a water-soluble contrast agent (diatrizoate meglumine and diatrizoate sodium [Gastroview, Mallinckrodt, St. Louis, MO]), and spot images were obtained using digital fluoroscopy equipment (Diagnost 76 Plus, Philips, Eindhoven, The Netherlands). If the spot images revealed a postoperative leak, the study was terminated without administration of barium. If the initial spot images failed to depict a leak, however, the patient was then given a 250% w/v barium suspension (E-Z-HD, E-Z-EM, Westbury, NY), and additional spot images were obtained. All of the studies were performed by residents, fellows, or attending physicians in gastrointestinal radiology, and all were interpreted by the attending radiologists.
In all 46 cases, the images were reviewed retrospectively by two experienced gastrointestinal radiologists to determine the location, morphology (i.e., blind-ending tracks, sealed-off collections, or free extravasation of contrast material), and size of the postoperative leaks. The original radiographic reports and images were reviewed retrospectively to determine whether the radiographic studies were performed with a water-soluble contrast agent alone or with a water-soluble contrast agent and high-density barium. In all cases in which both a water-soluble contrast agent and high-density barium were used, the images were also reviewed to determine whether, in retrospect, the leaks could be seen on the images from the initial phase of the examination with water-soluble contrast media. This assessment led to the reclassification of four cases as having been depicted on the images obtained with the water-soluble contrast agent despite subsequent administration of barium. The data were analyzed using a Fisher's exact test and Wilcoxon's rank sum test. The images from these 46 cases were also reviewed for the presence or absence of tracheobronchial aspiration. Finally, data were collected about follow-up radiographic studies of these patients.
Medical records were available for 33 (72%) of the 46 patients. The records were reviewed to determine the indications for surgery and whether the radiographic studies were obtained as routine tests to rule out postoperative leaks before resumption of oral feeding or whether the studies were obtained because of clinical suspicion of a leak. Finally, medical records were reviewed to determine whether detection of the leaks on radiographic studies affected patient management.
Our institutional review board approved all aspects of this retrospective study and did not require informed consent from patients whose records were included in our study.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Of the 33 patients with clinical follow-up, 18 (55%) had routine postoperative radiographic studies to rule out a leak before resumption of oral feeding and 15 (45%) had radiographic studies because of a clinically suspected leak. Of these 15 patients, 10 had fever, leukocytosis, or both; three had purulent drainage from their surgical incisions; one had a probable abscess on CT; and one had chest pain.
Radiographic Findings
Of the 46 leaks seen on radiographic studies after esophagogastrectomy,
total gastrectomy, or total laryngectomy, 23 (50%) were detected with a
water-soluble contrast agent alone and 23 (50%) were detected only after
additional administration of high-density barium (Figs.
1A,
1B,
2A,
2B,
3A,
3B,
4A,
4B). When the leaks were
classified on the basis of the type of surgery, the leaks were detected only
with high-density barium in 16 (67%) of 24 patients who underwent
esophagogastrectomy (Fig. 1A,
1B), in three (33%) of nine who
underwent total gastrectomy (Figs.
2A,
2B and
3A,
3B), and in four (31%) of 13
who underwent total laryngectomy (Fig.
4A,
4B). When the leaks were
classified on the basis of the indications for the radiographic study in the
33 patients with clinical follow-up, the leaks were detected only with
high-density barium in 15 (83%) of 18 patients who had routine studies before
resumption of oral feeding and in four (27%) of 15 who had studies because of
a clinically suspected leak.
|
|
|
|
|
|
|
|
Of the 23 leaks visualized on images obtained with water-soluble contrast media, six (26%) were characterized by blind-ending tracks, 14 (61%) by sealed-off collections, and three (13%) by free extravasation of contrast material into the mediastinum or neck. Of the 23 leaks visualized only after additional administration of high-density barium, 19 (83%) were characterized by blind-ending tracks (Figs. 3A, 3B and 4A, 4B) and four (17%) by sealed-off collections (Figs. 1A, 1B and 2A, 2B). The six tracks detected with a water-soluble contrast agent had a mean length of 2.0 cm (median length, 1.5 cm; range, 0.5–5.0 cm), whereas the 19 tracks detected with high-density barium had a mean length of 1.2 cm (median length, 1.0 cm; range, 0.3–3.0 cm). The 14 collections detected with a water-soluble contrast agent had a mean diameter of 4.6 cm (median diameter, 5.0 cm; range, 1.5–8.0 cm), whereas the four collections detected only with high-density barium had a mean diameter of 1.9 cm (median diameter, 1.3 cm; range, 1.0–3.0 cm). Thus, leaks detected with a water-soluble contrast agent were more likely to be characterized by discrete collections or free extravasation of contrast material than those detected with high-density barium, and the collections had a significantly larger mean diameter (p = 0.02). Conversely, leaks detected only with high-density barium were significantly more likely to be characterized by blind-ending tracks than those detected with a water-soluble contrast agent (p = 0.0007), and the tracks had a smaller mean length, although this difference was not found to be significant (p = 0.25).
Nineteen (79%) of the 24 leaks detected after esophagogastrectomy originated from the esophagogastric anastomosis (Fig. 1A, 1B) and the remaining five (21%) from the stapled portion of the intrathoracic stomach. All nine leaks (100%) detected after total gastrectomy originated from the esophagojejunal anastomosis (Figs. 2A, 2B and 3A, 3B). Finally, 12 (92%) of the 13 leaks detected after total laryngectomy originated from the surgically reconstructed pharynx (Fig. 4A, 4B) and one (8%) from a distal jejunal flap. Two patients aspirated contrast material into the tracheobronchial tree: one aspirated high-density barium, and the other aspirated water-soluble contrast media. Neither of these patients had symptoms or clinical complications attributable to aspiration of contrast material.
Follow-Up Radiographic Studies
Twenty-six patients had a total of 37 follow-up radiographic studies to
assess healing of the leaks, including one follow-up study in 16 patients, two
in nine patients, and three in one patient (mean, 1.4 studies; range,
1–3 studies). Fifteen (58%) of these 26 patients had persistent leaks on
the initial follow-up study, and 11 (42%) had no leaks. Of the 15 patients
with leaks, six had one or more additional radiographic studies that showed
persistent leaks, four had additional studies that showed no leaks, and five
had no other follow-up studies. In all, 22 (59%) of the 37 follow-up
radiographic studies showed persistent leaks and 15 (41%) showed no leaks.
Of the 22 leaks seen on follow-up radiographic studies, 11 (50%) were detected with a water-soluble contrast agent alone and 11 (50%) were detected only after additional administration of high-density barium. Of the 11 leaks visualized with water-soluble contrast media, seven (64%) were characterized by blind-ending tracks, three (27%) by sealed-off collections, and one (9%) by free extravasation of contrast material into the neck. Of the 11 leaks visualized only after additional administration of high-density barium, nine (82%) were characterized by blind-ending tracks and two (18%) by sealed-off collections. Finally, residual extraluminal barium in the neck or mediastinum was detected on the scout image in nine (24%) of the 37 follow-up studies (including five that showed persistent leaks and four that showed no leaks), but the residual barium did not compromise our ability to detect or rule out leaks in any of these cases.
Effect of Radiographic Studies on Clinical Management
The findings on postoperative radiographic studies affected management in
22 (67%) of the 33 patients with clinical follow-up. When the radiographic
studies were stratified on the basis of the choice of contrast agent, the
imaging findings affected management in 12 (86%) of 14 patients with leaks
detected on images obtained with water-soluble contrast media. Management
changes in these 12 patients included withholding oral feeding in six, delayed
removal of drainage catheters in three, placement of a new drainage catheter
in one, and repeated surgery to repair the leak in two. In 10 (53%) of the 19
patients with leaks detected only on images obtained with high-density barium,
management changes included withholding oral feeding in eight patients and
delayed removal of drainage catheters in two. Finally, when the leaks
affecting management were stratified on the basis of their morphology, nine
(75%) of the 12 leaks detected with a water-soluble contrast agent were
sealed-off collections, two (17%) were free leaks, and one (8%) was a
blind-ending track. Conversely, seven (70%) of the 10 leaks detected only with
high-density barium were blind-ending tracks, three (30%) were small
sealed-off collections, and none were free leaks. Thus, patient management was
altered by the detection of blind-ending tracks and sealed-off collections
with high-density barium that were not visualized on images obtained with a
water-soluble contrast agent.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Water-soluble contrast agents such as Gastroview are generally administered as the initial contrast medium for the detection of esophageal perforation because of a small theoretic risk that extravasated barium in the mediastinum may cause a granulomatous reaction with mediastinitis or mediastinal fibrosis, as previously shown in studies on laboratory cats [13, 14]. Water-soluble contrast agents also are absorbed rapidly from the mediastinum, so follow-up radiographic studies to assess the status of a leak are not compromised by residual contrast material in the mediastinum [1–4]. Nevertheless, water-soluble contrast agents are less radiopaque than barium and, when a leak is present, tend to disperse quickly in the mediastinum. As a result, radiographic studies performed with water-soluble contrast agents alone may fail to depict a substantial number of esophageal perforations [1, 5–7].
Because of the limitations of water-soluble contrast agents, most authors recommend that the radiographic examination be repeated immediately with barium to detect leaks, particularly subtle leaks, that are more likely to be visualized with a high-density contrast agent [1–4]. In studies by Tanomkiat and Galassi [8] and Buecker et al. [9], 38% and 22% of pharyngeal or esophageal perforations not visualized with water-soluble contrast agents subsequently were detected after additional administration of 60% w/v and 100% w/v barium suspensions, respectively. The ability of barium to show leaks not visualized with water-soluble contrast agents has been attributed to its greater radiopacity, better mucosal coating, and increased adherence to sites of extraluminal leakage [1–4]. Because water-soluble contrast agents are hyperosmolar and can draw fluid into the lungs causing pulmonary edema if aspirated into the tracheobronchial tree [2–4], some investigators even recommend that barium be used as the initial contrast medium to rule out anastomotic leaks after esophagogastrectomy [15].
In our study, a 250% w/v barium suspension depicted 50% of leaks that were not visible with a water-soluble contrast agent in patients who had undergone esophagogastrectomy, total gastrectomy, or total laryngectomy (Figs. 1A, 1B, 2A, 2B, 3A, 3B, 4A, 4B). This 250% w/v barium depicted an even higher rate of leaks missed with water-soluble contrast agents than the 60% w/v barium and 100% w/v barium used in earlier studies [8, 9], presumably because of its greater density and greater adherence to sites of extraluminal leakage. Our findings therefore support the concept that high-density barium, such as that used for routine double-contrast upper gastrointestinal tract examinations, is more effective than low-density barium for detecting pharyngeal or esophageal perforations not visualized with the use of water-soluble contrast agents. Another alternative is to administer low-density barium after the water-soluble contrast agent, and if no leak is identified, repeat the study with a high-density barium; however, because of the retrospective nature of our investigation, we could not evaluate this approach.
We also found that leaks detected on images obtained with water-soluble contrast agents were more likely to be characterized by extraluminal collections or free extravasation of contrast material into the mediastinum or neck, whereas leaks detected only on images obtained with high-density barium were significantly more likely to be characterized by blind-ending tracks (p = 0.0007). Also, sealed-off collections detected only with high-density barium had a significantly smaller mean diameter than those detected with a water-soluble contrast agent (p = 0.02). It therefore could be argued that the smaller tracks and collections detected only on images obtained with high-density barium are less likely to be clinically important leaks that affect patient management. In our study, however, management was affected in the majority (53%) of patients with small, blind-ending tracks or collections detected only with high-density barium. The most common effects on management included continued withholding of oral feeding or delayed removal of drainage catheters to allow the leaks to heal. Thus, the detection of even subtle perforations with high-density barium altered the management of these individuals.
In another study by Keberle et al. [16], 17 patients had pharyngeal leaks detected with a water-soluble contrast agent, but subsequent administration of a 100% w/v barium suspension failed to reveal two of these leaks and did not lead to the detection of any additional leaks not visualized with the water-soluble contrast agent. The authors concluded that pharyngeal perforations are more likely to be shown with water-soluble contrast agents than with barium, possibly because the greater density and viscosity of barium prevent it from entering narrow fistulous tracks in the neck. In our study, however, pharyngeal perforations were detected only with high-density barium in four (31%) of the 13 patients who had follow-up radiographic examinations after undergoing total laryngectomy (Fig. 4A, 4B). Our data suggest that high-density barium should be used to detect leaks that are not visualized with water-soluble contrast agents not only after esophageal surgery but also after pharyngeal surgery.
Our investigation is limited by the inherent selection biases of a retrospective study. Because of the limitations of our computerized engine search, this study also was confined to patients with postsurgical leaks, so we were unable to determine the usefulness of high-density barium for detection of leaks in patients with pharyngeal or esophageal perforations resulting from other causes. Finally, the limitations of our data collection process resulted in uncertainty about the timing of surgery in patients with leaks after esophagogastrectomy (i.e., some patients had radiographic studies in the remote postoperative period for other indications), so we were unable to determine the prevalence of postoperative leaks in our study population.
In summary, a 250% w/v barium suspension detected 50% of leaks that could not be visualized with a water-soluble contrast agent after esophagogastrectomy, total gastrectomy, or total laryngectomy. This high-density barium revealed a higher rate of leaks missed with water-soluble contrast agents than the 60% and 100% w/v bariums used in earlier studies [8, 9], presumably because of its greater density and greater adherence to sites of leakage. Although the leaks detected only with high-density barium were more likely to be characterized by small, blind-ending tracks than those visualized with a water-soluble contrast agent, patient management nevertheless was affected in the majority of cases. Our findings therefore support the use of high-density barium as part of the routine postoperative radiographic examination when no leaks are detected with a water-soluble contrast agent.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S UPPONI, A GANESHAN, H D'COSTA, M BETTS, N MAYNARD, H BUNGAY, and A SLATER Radiological detection of post-oesophagectomy anastomotic leak -- a comparison between multidetector CT and fluoroscopy Br. J. Radiol., July 1, 2008; 81(967): 545 - 548. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Panebianco, F. Francioni, M. Anzidei, M. Anile, M. Rolla, and R. Passariello Magnetic resonance-fluoroscopy as long-term follow-up examination in patients with narrow gastric tube reconstruction after radical esophagectomy. Eur. J. Cardiothorac. Surg., October 1, 2006; 30(4): 663 - 668. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ercan, T. W. Rice, S. C. Murthy, L. A. Rybicki, and E. H. Blackstone Does esophagogastric anastomotic technique influence the outcome of patients with esophageal cancer? J. Thorac. Cardiovasc. Surg., March 1, 2005; 129(3): 623 - 631. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Fadoo, D. E. Ruiz, S. K. Dawn, W. R. Webb, and M. B. Gotway Helical CT Esophagography for the Evaluation of Suspected Esophageal Perforation or Rupture Am. J. Roentgenol., May 1, 2004; 182(5): 1177 - 1179. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
