- Email: [email protected]
Management and long-term prognosis of Dieulafoy lesion
Management and long-term prognosis of Dieulafoy lesion Ian D. Norton, MB, BS, Bret T. Petersen, MD, Darius Sorbi, MD, Rita K. Balm, Glenn L. Alexander, Christopher J. Gostout, MD Rochester, Minnesota
Background: The Dieulafoy lesion is an important cause of gastrointestinal (GI) hemorrhage. Optimal treatment and long-term outcome are unknown. This study aimed to characterize the presentation of the Dieulafoy lesion and to summarize the results and report the long-term outcome of endoscopic therapy. Methods: Data regarding diagnosis, treatment and outcomes were derived from our GI Bleed Team database, patient records and follow-up correspondence. Results: Ninety Dieulafoy lesions were identified in 89 patients after a mean of 1.9 endoscopies. Their mean age was 72 years. Thirty-four percent of lesions were extragastric. Median transfusion requirement was 5 units. Two patients exsanguinated and 3 required surgery; all others were initially successfully treated endoscopically (with or without epinephrine injection): heat probe (71 patients), band ligation (3), hemoclip (1), laser (2), bipolar probe (4), sclerotherapy (2) and epinephrine alone (2). Gastric perforation occurred in 1 patient following sclerotherapy. Thirty-day mortality was 13%, 4 related to hemorrhage and 5 related to comorbidity. During median follow-up of 17 months, 34 patients (42%) died. One patient had recurrent bleeding 6 years after operation. Conclusions: Dieulafoy lesion is relatively common and often extragastric. Endoscopic therapy is safe and effective. Long-term recurrence was not evident following endoscopic ablation. Followup after ablative therapy appears unnecessary. (Gastrointest Endosc 1999;50:762-7.)
Dieulafoy lesion (exulceratio simplex, cirsoid aneurysm, caliber-persistent vessel) is an important cause of GI hemorrhage, accounting for up to 5% of acute upper GI bleeding.1-4 Although first reported by Gallard5 in 1896, the lesion was more accurately described by Dieulafoy6 in 1897. The histologic appearance is characteristic: a relatively large artery lies in close proximity to the mucosal surface, likely as a congenital anomaly.7,8 Currently, Dieulafoy lesions are usually obliterated using endoscopic methods2,4,9-11 and therefore tissue is rarely available for histologic diagnosis. The pathogenesis of vessel rupture and hemorrhage is uncertain but may be due to necrosis of the vessel wall induced by chronic gastritis.8 Hemorrhage is often torrential and may be life threatening. Previous descriptions of the Dieulafoy lesion emphasized a predilection for the proximal stomach. However, it is increasingly apparent that these lesions, although usually gastric, may develop throughout the GI tract.12-21 In this report the term Dieulafoy lesion is used to describe a lesion with Received November 30, 1998. For revision January 21, 1999. Accepted April 29, 1999. From the Department of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota. Reprint requests: Bret T. Petersen, MD, Division of Gastroenterology and Hepatology, Mayo Clinic, 200 1st St., SW, Rochester, MN 55905. Copyright © 1999 by the American Society for Gastrointestinal Endoscopy 0016-5107/99/$8.00 + 0 37/1/99624 762
GASTROINTESTINAL ENDOSCOPY
consistent clinical and endoscopic features present anywhere throughout the GI tract. Despite the importance of the Dieulafoy lesion in clinical practice, there is a paucity of information available regarding the long-term outcome of endoscopic treatment. The aims of this study were to determine the clinical presentation, efficacy of endoscopic therapy, and long-term prognosis of the Dieulafoy lesion in a large cohort of patients. PATIENTS AND METHODS Patients with a diagnosis of Dieulafoy lesion (Fig. 1) between October 1988 and July 1997 were identified from a comprehensive GI hemorrhage database accrued by the GI Bleed Team of our institution. The database contains comprehensive information on each patient, which accrues on first contact with the patient by GI Bleed Team members and continues throughout the patient’s hospitalization. Data are corroborated by the nurse coordinator and staff consultant on a daily basis. The database contains the following information: age, gender, comorbidities, medications, presenting symptoms and findings, transfusion types and quantities, all diagnostic and therapeutic interventions (endoscopic, radiologic, surgical), key endoscopic findings, detail of endoscopic therapy, length of stay (including intensive care unit), complications, information on recurrence of bleeding, presumptive and final diagnoses and mortality (related and unrelated). The design of our GI Bleed Team has previously been reported.22 In brief, it comprises 7 staff gastroenterologists with a specific interest in therapeutic endoscopy, a GI (or Advanced Endoscopy) Fellow on rotation through VOLUME 50, NO. 6, 1999
Management and long-term prognosis of Dieulafoy lesion
I Norton, B Petersen, D Sorbi, et al.
Figure 1. Fundal Dieulafoy lesion.
Figure 2. Fundal Dieulafoy lesion after combination therapy with epinephrine injection and heat probe thermocoagulation.
this service and a nurse coordinator who collates and enters the data. All procedures were performed in the presence of a staff member of the GI Bleed Team. Patients with evidence of ongoing bleeding underwent endoscopy either immediately or, if necessary, after a brief period of resuscitation (median time to endoscopy 1.5 hours). Endoscopy was performed electively on nonbleeding patients within 24 hours of presentation. All patients with clinically evident recurrent bleeding underwent prompt assessment and re-endoscopy. Colonoscopy (when indicated) was typically performed following a rapid bowel lavage using 3 to 4 L polyethylene glycol electrolyte solution administered via nasogastric tube during a period of 2 hours. All endoscopies were performed using Olympus videoendoscopes (Olympus America, Inc., Melville, N.Y.). Typically, the procedure was commenced with a standard (GIF 100) videoendoscope. In the event of excessive blood obscuring the endoscopic view, this instrument was removed and the XGIF-WT2 (6 mm accessory channel) fiberoptic endoscope was used to clear the field before resuming with this or a standard diagnostic instrument. In the event that upper endoscopy did not reveal a bleeding site or obvious features of an upper GI lesion (such as a large amount of blood in the stomach), the patient underwent colonoscopy. Blood seen in the duodenum or terminal ileum, or the absence of lesions in the stomach and colon, led to enteroscopy with either the PCF-100 pediatric colonoscope or the SIF-100 push enteroscope. The endoscopic diagnosis of Dieulafoy lesion was based on the following established criteria14: (1) active arterial spurting or micropulsatile streaming from a minute (less than 3 mm) mucosal defect; (2) visualization of a protruding vessel with or without active bleeding within a minute mucosal defect with normal surrounding mucosa; or (3) densely adherent clot with a narrow point of attachment to a minute mucosal defect or normal appearing mucosa. To avoid possible inclusion of variceal bleeding at either index or recurrent bleeding, all patients with esophagogastric varices were excluded from the study.
Table 1. Patient characteristics
VOLUME 50, NO. 6, 1999
Number Age, median (yr) (range) Gender (M/F) Medications (%) Aspirin/NSAIDs Coumadin Comorbidity (%) Cardiovascular Hypertension Diabetes Malignancy Chronic liver disease Renal disease Inpatient status
90 72 (30-94) 53:37 51 13 90 60 44 17 13 11 9 11%
Endoscopic therapy was administered in 85 cases. Epinephrine was usually, but not universally, used prior to more definitive therapy. Ablative treatments included heat probe (71 patients, Fig. 2), band ligation (3 patients, Fig. 3), hemoclip (1), laser (2), bipolar probe (4), sclerotherapy (2) and epinephrine alone (2). Information regarding diagnosis, treatment and outcomes was derived from the database and supplemented with a review of patient records. Further patient follow-up was via a questionnaire mailed to all patients or a relative. In the event that the patient could not be contacted, follow-up was derived from the last appointment containing specific reference to the patient’s bleeding history. Only evidence of recurrent acute GI bleeding was sought; occult blood loss was not interpreted as recurrent bleeding from a Dieulafoy lesion. Descriptive statistics were derived using the Excel spreadsheet program (Microsoft Corp., Redmond, Wash.). The study was reviewed and approved by our institutional review board. In accordance with state laws, only patients consenting to inclusion in database or medical record based research protocols were included in the study. One patient denied consent for participation in this study. GASTROINTESTINAL ENDOSCOPY
763
I Norton, B Petersen, D Sorbi, et al.
Figure 3. Fundal Dieulafoy lesion (translucent visible vessel) after band ligation.
RESULTS Ninety-five Dieulafoy lesions were identified in 94 patients during the study period. Five patients whose bleeding had been attributed to a Dieulafoy lesion in the presence of varices were excluded from the study, leaving 90 Dieulafoy lesions in 89 patients for analysis. During this period the GI Bleed Team was consulted on 4804 episodes of acute GI bleeding. Therefore, Dieulafoy lesion accounted for 1.9% of all GI Bleed Team consultations for acute GI hemorrhage. This percentage was relatively constant during the study period in spite of an increase in the number of consultations performed by this service. Detailed information concerning the patients is given in Table 1. Fifty-three patients (60%) were men. Median age at presentation was 72 years (range 30 to 94 years). Eleven percent were hospitalized at the time of hemorrhage. Significant comorbidity was present in 90%, most commonly ischemic heart disease (62%) and hypertension (44%). Cirrhosis had been previously diagnosed on biopsy or radiologic grounds in 8 patients (9%), but none of these patients had gastroesophageal varices at endoscopy. This is likely to be an underestimate of the prevalence of cirrhosis among patients with Dieulafoy lesion, because we excluded patients with varices. Fifty-five patients (61%) were taking medications that adversely affect coagulation: 51% were taking aspirin or nonsteroidal anti-inflammatory drugs (NSAIDs) and 13% warfarin. All patients presented with acute GI hemorrhage. Of particular interest, hematochezia was the presenting symptom in 28 (31%) patients overall and 19 of 81 (23%) with a lesion proximal to the ligament of Treitz. Sixty-nine lesions (77%) were bleeding at the time of diagnostic endoscopy, 13 (14%) had adherent clot without active bleeding and 8 (9%) 764
GASTROINTESTINAL ENDOSCOPY
Management and long-term prognosis of Dieulafoy lesion
Figure 4. Pie chart depicting the sites of Dieulafoy lesions.
demonstrated a nonbleeding vessel. Other identified lesions potentially responsible for GI hemorrhage included gastric erosions in 18% and peptic ulceration in 11% of the patients with a gastric or duodenal Dieulafoy lesion. Of the 9 patients with a colonic Dieulafoy lesion, 2 had significant diverticulosis. The initial endoscopy was diagnostic in 63% of patients. The mean number of endoscopies required for diagnosis was 1.9 (range 1 to 4). Failure to diagnose when the appropriate site was evaluated at endoscopy was due to excessive blood on 23 occasions (44%) and a missed lesion on 30 occasions (56%). The sites of Dieulafoy lesions are shown in Figure 4. Of note, one third were extragastric. Various endoscopic modalities and combinations were utilized to control bleeding. In 3 cases endoscopic management failed to control hemorrhage and urgent surgery was undertaken with performance of a hemicolectomy, oversew of a fundal lesion and resection of a fundal lesion. In all other cases hemostasis was achieved with endoscopic therapy. Eight patients (9%) had recurrent bleeding following initial therapy during the same hospital admission. All underwent emergency endoscopy and retreatment that resulted in permanent hemostasis in 7. One patient had recurrent bleeding twice prior to permanent hemostasis. No patient required surgery for recurrence of bleeding. The median transfusion requirement was 5 units of red blood cells (range 0 to 24 units). Median follow-up was 19 months (range 0 to 103 months), with 9 patients lost to follow-up since discharge from the hospital. Only 1 patient had recurrent bleeding from a Dieulafoy lesion subsequent to the initial hospitalization. This patient initially underwent oversewing of a fundal Dieulafoy (he did not undergo endoscopic therapy because the lesion was obscured by profuse bleeding). This patient bled VOLUME 50, NO. 6, 1999
Management and long-term prognosis of Dieulafoy lesion
6 years later in the same region of the stomach. Permanent hemostasis on this occasion was achieved with epinephrine injection and heat probe. He had no further bleeding but died of congestive heart failure 4 months later. Thirty-day mortality was 13%. Death was directly attributable to the bleeding episode in 4 patients: 1 patient exsanguinated on arrival in the emergency room, 1 patient in whom advanced malignancy precluded surgery exsanguinated after a failed attempt at endoscopic therapy and 2 patients died of cardiovascular causes within 48 hours of presentation. Thirty-four patients (42%) died during follow-up. Following hospital discharge, no deaths were attributable to the Dieulafoy lesion. DISCUSSION Dieulafoy lesion is likely an underdiagnosed rather than a rare disorder. It accounts for between 0.3% and 6.8% of acute GI hemorrhage.1,2,4,23 In our study the Dieulafoy lesion accounted for 1.9% of acute GI hemorrhage. Histologic confirmation of the diagnosis was available in only 1 patient. The patient demographics of our study are in accord with those of other published series. A review of 149 published cases reported a preponderance in men23 as did a recent report.2 Our patients were older (median age 72 versus mean ages 58 and 61 years, respectively, in the two aforementioned studies). However, all series confirm the wide age distribution, with patient age ranging from the third or fourth decade to the tenth decade. Excessive alcohol consumption has been reported to be an association of Dieulafoy lesion.23 Eight patients in our study (9%) had known cirrhosis, but we have insufficient data regarding alcohol consumption in the other patients to confirm or refute this relationship. This incidence of cirrhosis is likely to be an underestimate, because we excluded patients with esophagogastric varices from the study. As seen in a previous study,2 many patients in this series (90%) had significant comorbidity, in particular, cardiovascular disease, hypertension and diabetes. This is likely a reflection of the elderly population studied and any pathogenetic significance remains speculative. Pathologic studies have failed to demonstrate atherosclerosis or aneurysmal dilatation of the involved vessel.1,7 Ingestion of aspirin and NSAIDs was common in this study (51%), similar to that described by Baettig et al.2 (43%). This is not surprising in this elderly group of patients, many with cardiovascular disease. Juler et al.7 have postulated NSAID-induced gastritis with subsequent vessel wall necrosis in the pathogenesis of the Dieulafoy lesion rupture. However, in the current study the VOLUME 50, NO. 6, 1999
I Norton, B Petersen, D Sorbi, et al.
adjacent mucosa in all cases was normal at endoscopic assessment. Initial endoscopy was diagnostic in 63% of cases. This is in part due to the fact that patients presenting with hematochezia routinely undergo upper endoscopy in the first instance to exclude a proximal lesion prior to lavage for colonoscopy. Thus, all colonic Dieulafoy lesions required at least 2 endoscopic procedures for diagnosis. Excluding the initial upper GI endoscopy in patients with colonic lesions, the first endoscopy was diagnostic in 73%. A previous study has reported a sensitivity of 49% for initial endoscopy.23 In view of the intermittent nature of bleeding from the Dieulafoy lesion, the sensitivity of endoscopic diagnosis is likely to be increased through early endoscopy in patients with acute GI bleeding. One third of Dieulafoy lesions in this study were extragastric. Although there are numerous published descriptions of lesions in nongastric sites,12-21 no other series has reported this high proportion of extragastric Dieulafoy lesions. We suspect that the traditional association of Dieulafoy lesions with the proximal stomach yields a reluctance to use the term for lesions located at extragastric sites. It is likely that these are often classified as ulcer disease in the duodenum or vascular malformation in the duodenum or other sites. Endoscopic management in this series was highly successful, with primary hemostasis achieved in 96%. This is identical to the rate of primary hemostasis reported by Baettig et al.2 The only endoscopic complication in this study was gastric perforation following injection of sodium tetradecyl sulfate. We no longer use sclerosant injection for treatment of these lesions, preferring epinephrine injection followed by either band ligation or thermal ablation using either heat probe or bipolar probe. Endoscopic ablative therapy has been advocated in a recent review of the Dieulafoy lesion, although no randomized controlled trials have been performed.1 Many reports have described successful hemostasis utilizing a variety of endoscopic modalities, such as injection with sclerosant2,3 or cyanoacrylate glue,24 thermal ablation with heat probe,4,25 neodymiumyttrium aluminum garnet laser,26 monopolar27,28 or bipolar probe,4,9 and band ligation.10,29 Most experience is with sclerosant injection and thermal ablation. Contact thermal ablation (e.g., heat probe) has been shown in this and other studies to be the mainstay of management. The prior use of epinephrine injection may be a useful adjunct to slow or stop bleeding prior to ablative therapy or prior to clot guillotine. We do not advocate the use of epinephrine alone in the management of suspected Dieulafoy GASTROINTESTINAL ENDOSCOPY
765
I Norton, B Petersen, D Sorbi, et al.
lesion.4,25,27,28 Use of neodymium-yttrium aluminum garnet laser in our study was early in the course of the study and is no longer used due to its cumbersome nature.23 Nonthermal treatment such as hemoclip may play a role in patients with bleeding diathesis, where minimization of tissue damage is desirable. No study has compared the results of surgical and endoscopic therapy for Dieulafoy lesion. A randomized study comparing endoscopic and surgical modalities would be impractical, given the relative infrequency of Dieulafoy lesions and the success of endoscopic therapies in uncontrolled series. This study demonstrates that when endoscopic therapy is unsuccessful the endoscopic examination still has an important role for identifying the lesion, because a nonbleeding Dieulafoy lesion may be undetectable at gastrotomy. This has led some authors to advocate a combined endoscopic and surgical approach to the patient requiring operation.30,31 Furthermore, if obliterative endoscopic therapy is not undertaken or is unsuccessful, it may be useful to tattoo the site of the bleeding lesion to guide the surgeon. Nine percent of subjects had further bleeding during the same hospital stay following endoscopic therapy. However, this was controlled at repeat endoscopy in all cases. This is comparable to the rates of recurrent bleeding in other large series including 11% in the series of Asaki et al.3 (46 patients), 11% in that of Baettig et al.2 (28 patients) and 22% in the series of Pointner et al.4 (11 patients). A review by Reilly and Al-Kawas23 of 177 Dieulafoy lesions quoted an overall rate of recurrent bleeding of 15%. Surgical intervention was required in 3 patients in whom bleeding could not be controlled endoscopically. This is consistent with three recent studies in which the rate of surgical intervention was between 2% and 6%.2-4 Angiography with Gelfoam embolization was reported to successfully arrest hemorrhage in 3 of 4 patients in one study.23 None of the patients in our cohort underwent angiographic therapy. It is likely that this modality may be particularly useful when endoscopic therapy has failed and the patient is deemed a poor candidate for surgery. Thirty-day mortality in this study was high (12 patients, 13%). Four early deaths were directly attributable to hemorrhage or its consequences. Eight additional deaths occurred within 30 days of the index bleeding episode but were not thought to be related to the hemorrhage or its consequences. During the follow-up period (median of 19 months), mortality was high (42%). However, no further death was directly related to the index episode 766
GASTROINTESTINAL ENDOSCOPY
Management and long-term prognosis of Dieulafoy lesion
or its sequelae. Similar rates were noted in another study, with 12% in-hospital and 45% overall mortality noted during a 14-month follow-up period.2 These investigators also believed that the mortality was largely a reflection of patient comorbidity. Due to the relative infrequency of this diagnosis, few large series of patients with extended follow-up have been reported. In this study, following definitive therapy and discharge from hospital, only 1 patient suffered a relapse of bleeding from a Dieulafoy lesion. This was one of the few patients who did not undergo “obliterative” therapy by either endoscopic or surgical means. The patient in question had an operative oversew of the original site and bled at the same site 6 years after the index episode. In summary, Dieulafoy lesion appears to account for approximately 2% of acute GI hemorrhage. Endoscopic therapy is effective and well tolerated and is of long-term benefit. No patients in our series have bled from more than 1 Dieulafoy lesion. Thirtyday and subsequent mortality following hemorrhage from a Dieulafoy lesion is high but relates almost exclusively to the patient’s comorbid status. ACKNOWLEDGMENTS The other members, both past and present, of our GI Bleed Team (1988-1999) are: Drs. David A. Ahlquist, Jeffrey A. Alexander, Todd H. Baron, Jonathan E. Clain, Rollin W. Hughes, Patrick S. Kamath, Mark V. Larson, Edward V. Lofthus, Darrell S. Pardi, John J. Poterucha, Kenneth W. Schroeder, William J. Tremaine, Thomas R. Viggiano, and Kenneth K. Wang. REFERENCES 1. Katz PO, Salas L. Less frequent causes of upper gastrointestinal bleeding [review]. Gastroenterol Clin North Am 1993;22:875-89. 2. Baettig B, Haecki W, Lammer F, Jost R. Dieulafoy’s disease: endoscopic treatment and follow up. Gut 1993;34:1418-21. 3. Asaki S, Sato H, Nishimura T, Ohkubo S, Yamagata R, Ito S, et al. Endoscopic diagnosis and treatment of Dieulafoy’s ulcer. Tohuko J Exp Med 1988;154:135-41. 4. Pointner R, Schwab G, Konigsrainer A, Dietze O. Endoscopic treatment of Dieulafoy’s disease. Gastroenterology 1988;94: 563-6. 5. Gallard T. Miliary aneurisms of the stomach giving cause to fatal hematemesis. Bull Soc Med Hop Paris 1894;1:84-91. 6. Dieulafoy G. Exulceration Simplex: L’intervention chirurgicale dans la hematemeses foudroyantes consecutives a l’exulceration simple de l’esomac. Bull Acad Med 1898;39:49-84. 7. Juler GL, Labitzke HG, Lamb R, Allen R. The pathogenesis of Dieulafoy’s gastric erosion. Am J Gastroenterol 1984;79:195200. 8. Molnar P, Miko T. Multiple arterial caliber-persistence resulting in hematomas and fatal rupture of the gastric wall. Am J Surg Pathol 1982;6:83-6. 9. Boron B, Mobarham S. Endoscopic treatment of Dieulafoy hemorrhage. J Clin Gastroenterol 1987;9:518-20. VOLUME 50, NO. 6, 1999
Management and long-term prognosis of Dieulafoy lesion
10. Brown GR, Harford WV, Jones WF. Endoscopic band ligation of an actively bleeding Dieulafoy lesion. Gastrointest Endosc 1994;40:501-3. 11. Parra-Blanco A, Takahashi H, Mendez Jerez PV, Kojima T, Aksoz K, Kirihara K, et al. Endoscopic management of Dieulafoy lesions of the stomach: a case study of 26 patients. Endoscopy 1997;29:834-9. 12. Bejanin H, Boivin C, Dehni N, Pipien I, Bloch F, Bruneval P, et al. Dieulafoy hemorrhagic duodenal ulcer: first case histologically confirmed [letter]. Gastroenterol Clin et Biol 1995; 19:227-8. 13. Barbier P, Luder P, Triller J, Ruchti C, Hassler Stafford A. Colonic hemorrhage from a solitary minute ulcer: report of three cases. Gastroenterology 1985;88:1065-8. 14. Dy NM, Gostout CJ, Balm RK. Bleeding from the endoscopically-identified Dieulafoy lesion of the proximal small intestine and colon. Am J Gastroenterol 1995;90:108-11. 15. Eguchi S, Maeda J, Taguchi H, Kanematsu T. Massive gastrointestinal bleeding from a Dieulafoy-like lesion of the rectum. J Clin Gastroenterol 1997;24:262-3. 16. Goins WA, Chatman DM, Kaviani MJ. Massive lower gastrointestinal bleeding due to “Dieulafoy’s vascular malformation” of the jejunum: case report. J Natl Med Assoc 1995;87: 766-70. 17. Goldenberg SP, DeLuca VA Jr, Marignani P. Endoscopic treatment of Dieulafoy’s lesion of the duodenum. Am J Gastroenterol 1990;85:452-2. 18. Paraf F. Dieulafoy disease of the gallbladder: first case? [letter]. Gastroenterol Clin et Biol 1996;20:712. 19. Wetscher G, Schwab G, Glaser K, Fend F, Bodner E, Pointner R. Dieulafoy lesion in a congenital double pylorus [letter]. Endoscopy 1994;26:374-5. 20. Scheider DM, Barthel JS, King PD, Beale GD. Dieulafoy-like
VOLUME 50, NO. 6, 1999
I Norton, B Petersen, D Sorbi, et al.
21.
22.
23. 24.
25.
26. 27. 28. 29.
30.
31.
lesion of the distal esophagus. Am J Gastroenterol 1994;89: 2080-1. Richards WO, Grove-Mahoney D, Williams LF. Hemorrhage from a Dieulafoy type ulcer of the colon: a new cause of lower gastrointestinal bleeding. Am Surg 1988;54:121-4. Gostout CJ, Wang KK, Ahlquist DL, Clain JE, Hughes RW, Larson MV, et al. Acute gastrointestinal bleeding: experience of a specialized management team. J Clin Gastroenterol 1992;14:260-7. Reilly HF, Al-Kawas FH. Dieulafoy’s lesion: diagnosis and management. Dig Dis Sci 1991;36:1702-7. D’Imperio D, Papadia C, Baroncini D, Piemontese A, Billi P. N-butyl-2-cyanoacrylate in the endoscopic treatment of Dieulafoy ulcer [abstract]. Endoscopy 1995;27:216. Lin HJ, Lee FY, Tsai YT, Lee SD, Lee CH, Kang WM. Therapeutic endoscopy for Dieulafoy’s disease. J Clin Gastroenterol 1989;11:507-10. Al Kawas FH, O’Keefe J. Nd:YAG laser treatment of a bleeding Dieulafoy lesion. Gastrointest Endosc 1987;33:38-9. Bakka A, Rosseland AR. Massive gastric bleeding from exulceratio simplex (Dieulafoy). Acta Chir Scand 1986;152:285-8. Hoffman J, Beck H, Jensen HE. Dieulafoy’s lesion. Surg Gynecol Obstet 1984;159:537-40. Murray KF, Jennings RW, Fox VL. Endoscopic band ligation of a Dieulafoy lesion in the small intestine of a child. Gastrointest Endosc 1996;44:336-9. Grisendi A, Lonardo A, Della Casa G, Frazzoni M, Pulvirenti M, Ferrari AM, et al. Combined endoscopic and surgical management of Dieulafoy vascular malformation. J Am Coll Surg 1994;179:182-6. Mixter CG, Sullivan CA. Control of proximal gastric bleeding: combined laparoscopic and endoscopic approach. J Laparoendosc Surg 1992;2:105-9.
GASTROINTESTINAL ENDOSCOPY
767
Background: The Dieulafoy lesion is an important cause of gastrointestinal (GI) hemorrhage. Optimal treatment and long-term outcome are unknown. This study aimed to characterize the presentation of the Dieulafoy lesion and to summarize the results and report the long-term outcome of endoscopic therapy. Methods: Data regarding diagnosis, treatment and outcomes were derived from our GI Bleed Team database, patient records and follow-up correspondence. Results: Ninety Dieulafoy lesions were identified in 89 patients after a mean of 1.9 endoscopies. Their mean age was 72 years. Thirty-four percent of lesions were extragastric. Median transfusion requirement was 5 units. Two patients exsanguinated and 3 required surgery; all others were initially successfully treated endoscopically (with or without epinephrine injection): heat probe (71 patients), band ligation (3), hemoclip (1), laser (2), bipolar probe (4), sclerotherapy (2) and epinephrine alone (2). Gastric perforation occurred in 1 patient following sclerotherapy. Thirty-day mortality was 13%, 4 related to hemorrhage and 5 related to comorbidity. During median follow-up of 17 months, 34 patients (42%) died. One patient had recurrent bleeding 6 years after operation. Conclusions: Dieulafoy lesion is relatively common and often extragastric. Endoscopic therapy is safe and effective. Long-term recurrence was not evident following endoscopic ablation. Followup after ablative therapy appears unnecessary. (Gastrointest Endosc 1999;50:762-7.)
Dieulafoy lesion (exulceratio simplex, cirsoid aneurysm, caliber-persistent vessel) is an important cause of GI hemorrhage, accounting for up to 5% of acute upper GI bleeding.1-4 Although first reported by Gallard5 in 1896, the lesion was more accurately described by Dieulafoy6 in 1897. The histologic appearance is characteristic: a relatively large artery lies in close proximity to the mucosal surface, likely as a congenital anomaly.7,8 Currently, Dieulafoy lesions are usually obliterated using endoscopic methods2,4,9-11 and therefore tissue is rarely available for histologic diagnosis. The pathogenesis of vessel rupture and hemorrhage is uncertain but may be due to necrosis of the vessel wall induced by chronic gastritis.8 Hemorrhage is often torrential and may be life threatening. Previous descriptions of the Dieulafoy lesion emphasized a predilection for the proximal stomach. However, it is increasingly apparent that these lesions, although usually gastric, may develop throughout the GI tract.12-21 In this report the term Dieulafoy lesion is used to describe a lesion with Received November 30, 1998. For revision January 21, 1999. Accepted April 29, 1999. From the Department of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota. Reprint requests: Bret T. Petersen, MD, Division of Gastroenterology and Hepatology, Mayo Clinic, 200 1st St., SW, Rochester, MN 55905. Copyright © 1999 by the American Society for Gastrointestinal Endoscopy 0016-5107/99/$8.00 + 0 37/1/99624 762
GASTROINTESTINAL ENDOSCOPY
consistent clinical and endoscopic features present anywhere throughout the GI tract. Despite the importance of the Dieulafoy lesion in clinical practice, there is a paucity of information available regarding the long-term outcome of endoscopic treatment. The aims of this study were to determine the clinical presentation, efficacy of endoscopic therapy, and long-term prognosis of the Dieulafoy lesion in a large cohort of patients. PATIENTS AND METHODS Patients with a diagnosis of Dieulafoy lesion (Fig. 1) between October 1988 and July 1997 were identified from a comprehensive GI hemorrhage database accrued by the GI Bleed Team of our institution. The database contains comprehensive information on each patient, which accrues on first contact with the patient by GI Bleed Team members and continues throughout the patient’s hospitalization. Data are corroborated by the nurse coordinator and staff consultant on a daily basis. The database contains the following information: age, gender, comorbidities, medications, presenting symptoms and findings, transfusion types and quantities, all diagnostic and therapeutic interventions (endoscopic, radiologic, surgical), key endoscopic findings, detail of endoscopic therapy, length of stay (including intensive care unit), complications, information on recurrence of bleeding, presumptive and final diagnoses and mortality (related and unrelated). The design of our GI Bleed Team has previously been reported.22 In brief, it comprises 7 staff gastroenterologists with a specific interest in therapeutic endoscopy, a GI (or Advanced Endoscopy) Fellow on rotation through VOLUME 50, NO. 6, 1999
Management and long-term prognosis of Dieulafoy lesion
I Norton, B Petersen, D Sorbi, et al.
Figure 1. Fundal Dieulafoy lesion.
Figure 2. Fundal Dieulafoy lesion after combination therapy with epinephrine injection and heat probe thermocoagulation.
this service and a nurse coordinator who collates and enters the data. All procedures were performed in the presence of a staff member of the GI Bleed Team. Patients with evidence of ongoing bleeding underwent endoscopy either immediately or, if necessary, after a brief period of resuscitation (median time to endoscopy 1.5 hours). Endoscopy was performed electively on nonbleeding patients within 24 hours of presentation. All patients with clinically evident recurrent bleeding underwent prompt assessment and re-endoscopy. Colonoscopy (when indicated) was typically performed following a rapid bowel lavage using 3 to 4 L polyethylene glycol electrolyte solution administered via nasogastric tube during a period of 2 hours. All endoscopies were performed using Olympus videoendoscopes (Olympus America, Inc., Melville, N.Y.). Typically, the procedure was commenced with a standard (GIF 100) videoendoscope. In the event of excessive blood obscuring the endoscopic view, this instrument was removed and the XGIF-WT2 (6 mm accessory channel) fiberoptic endoscope was used to clear the field before resuming with this or a standard diagnostic instrument. In the event that upper endoscopy did not reveal a bleeding site or obvious features of an upper GI lesion (such as a large amount of blood in the stomach), the patient underwent colonoscopy. Blood seen in the duodenum or terminal ileum, or the absence of lesions in the stomach and colon, led to enteroscopy with either the PCF-100 pediatric colonoscope or the SIF-100 push enteroscope. The endoscopic diagnosis of Dieulafoy lesion was based on the following established criteria14: (1) active arterial spurting or micropulsatile streaming from a minute (less than 3 mm) mucosal defect; (2) visualization of a protruding vessel with or without active bleeding within a minute mucosal defect with normal surrounding mucosa; or (3) densely adherent clot with a narrow point of attachment to a minute mucosal defect or normal appearing mucosa. To avoid possible inclusion of variceal bleeding at either index or recurrent bleeding, all patients with esophagogastric varices were excluded from the study.
Table 1. Patient characteristics
VOLUME 50, NO. 6, 1999
Number Age, median (yr) (range) Gender (M/F) Medications (%) Aspirin/NSAIDs Coumadin Comorbidity (%) Cardiovascular Hypertension Diabetes Malignancy Chronic liver disease Renal disease Inpatient status
90 72 (30-94) 53:37 51 13 90 60 44 17 13 11 9 11%
Endoscopic therapy was administered in 85 cases. Epinephrine was usually, but not universally, used prior to more definitive therapy. Ablative treatments included heat probe (71 patients, Fig. 2), band ligation (3 patients, Fig. 3), hemoclip (1), laser (2), bipolar probe (4), sclerotherapy (2) and epinephrine alone (2). Information regarding diagnosis, treatment and outcomes was derived from the database and supplemented with a review of patient records. Further patient follow-up was via a questionnaire mailed to all patients or a relative. In the event that the patient could not be contacted, follow-up was derived from the last appointment containing specific reference to the patient’s bleeding history. Only evidence of recurrent acute GI bleeding was sought; occult blood loss was not interpreted as recurrent bleeding from a Dieulafoy lesion. Descriptive statistics were derived using the Excel spreadsheet program (Microsoft Corp., Redmond, Wash.). The study was reviewed and approved by our institutional review board. In accordance with state laws, only patients consenting to inclusion in database or medical record based research protocols were included in the study. One patient denied consent for participation in this study. GASTROINTESTINAL ENDOSCOPY
763
I Norton, B Petersen, D Sorbi, et al.
Figure 3. Fundal Dieulafoy lesion (translucent visible vessel) after band ligation.
RESULTS Ninety-five Dieulafoy lesions were identified in 94 patients during the study period. Five patients whose bleeding had been attributed to a Dieulafoy lesion in the presence of varices were excluded from the study, leaving 90 Dieulafoy lesions in 89 patients for analysis. During this period the GI Bleed Team was consulted on 4804 episodes of acute GI bleeding. Therefore, Dieulafoy lesion accounted for 1.9% of all GI Bleed Team consultations for acute GI hemorrhage. This percentage was relatively constant during the study period in spite of an increase in the number of consultations performed by this service. Detailed information concerning the patients is given in Table 1. Fifty-three patients (60%) were men. Median age at presentation was 72 years (range 30 to 94 years). Eleven percent were hospitalized at the time of hemorrhage. Significant comorbidity was present in 90%, most commonly ischemic heart disease (62%) and hypertension (44%). Cirrhosis had been previously diagnosed on biopsy or radiologic grounds in 8 patients (9%), but none of these patients had gastroesophageal varices at endoscopy. This is likely to be an underestimate of the prevalence of cirrhosis among patients with Dieulafoy lesion, because we excluded patients with varices. Fifty-five patients (61%) were taking medications that adversely affect coagulation: 51% were taking aspirin or nonsteroidal anti-inflammatory drugs (NSAIDs) and 13% warfarin. All patients presented with acute GI hemorrhage. Of particular interest, hematochezia was the presenting symptom in 28 (31%) patients overall and 19 of 81 (23%) with a lesion proximal to the ligament of Treitz. Sixty-nine lesions (77%) were bleeding at the time of diagnostic endoscopy, 13 (14%) had adherent clot without active bleeding and 8 (9%) 764
GASTROINTESTINAL ENDOSCOPY
Management and long-term prognosis of Dieulafoy lesion
Figure 4. Pie chart depicting the sites of Dieulafoy lesions.
demonstrated a nonbleeding vessel. Other identified lesions potentially responsible for GI hemorrhage included gastric erosions in 18% and peptic ulceration in 11% of the patients with a gastric or duodenal Dieulafoy lesion. Of the 9 patients with a colonic Dieulafoy lesion, 2 had significant diverticulosis. The initial endoscopy was diagnostic in 63% of patients. The mean number of endoscopies required for diagnosis was 1.9 (range 1 to 4). Failure to diagnose when the appropriate site was evaluated at endoscopy was due to excessive blood on 23 occasions (44%) and a missed lesion on 30 occasions (56%). The sites of Dieulafoy lesions are shown in Figure 4. Of note, one third were extragastric. Various endoscopic modalities and combinations were utilized to control bleeding. In 3 cases endoscopic management failed to control hemorrhage and urgent surgery was undertaken with performance of a hemicolectomy, oversew of a fundal lesion and resection of a fundal lesion. In all other cases hemostasis was achieved with endoscopic therapy. Eight patients (9%) had recurrent bleeding following initial therapy during the same hospital admission. All underwent emergency endoscopy and retreatment that resulted in permanent hemostasis in 7. One patient had recurrent bleeding twice prior to permanent hemostasis. No patient required surgery for recurrence of bleeding. The median transfusion requirement was 5 units of red blood cells (range 0 to 24 units). Median follow-up was 19 months (range 0 to 103 months), with 9 patients lost to follow-up since discharge from the hospital. Only 1 patient had recurrent bleeding from a Dieulafoy lesion subsequent to the initial hospitalization. This patient initially underwent oversewing of a fundal Dieulafoy (he did not undergo endoscopic therapy because the lesion was obscured by profuse bleeding). This patient bled VOLUME 50, NO. 6, 1999
Management and long-term prognosis of Dieulafoy lesion
6 years later in the same region of the stomach. Permanent hemostasis on this occasion was achieved with epinephrine injection and heat probe. He had no further bleeding but died of congestive heart failure 4 months later. Thirty-day mortality was 13%. Death was directly attributable to the bleeding episode in 4 patients: 1 patient exsanguinated on arrival in the emergency room, 1 patient in whom advanced malignancy precluded surgery exsanguinated after a failed attempt at endoscopic therapy and 2 patients died of cardiovascular causes within 48 hours of presentation. Thirty-four patients (42%) died during follow-up. Following hospital discharge, no deaths were attributable to the Dieulafoy lesion. DISCUSSION Dieulafoy lesion is likely an underdiagnosed rather than a rare disorder. It accounts for between 0.3% and 6.8% of acute GI hemorrhage.1,2,4,23 In our study the Dieulafoy lesion accounted for 1.9% of acute GI hemorrhage. Histologic confirmation of the diagnosis was available in only 1 patient. The patient demographics of our study are in accord with those of other published series. A review of 149 published cases reported a preponderance in men23 as did a recent report.2 Our patients were older (median age 72 versus mean ages 58 and 61 years, respectively, in the two aforementioned studies). However, all series confirm the wide age distribution, with patient age ranging from the third or fourth decade to the tenth decade. Excessive alcohol consumption has been reported to be an association of Dieulafoy lesion.23 Eight patients in our study (9%) had known cirrhosis, but we have insufficient data regarding alcohol consumption in the other patients to confirm or refute this relationship. This incidence of cirrhosis is likely to be an underestimate, because we excluded patients with esophagogastric varices from the study. As seen in a previous study,2 many patients in this series (90%) had significant comorbidity, in particular, cardiovascular disease, hypertension and diabetes. This is likely a reflection of the elderly population studied and any pathogenetic significance remains speculative. Pathologic studies have failed to demonstrate atherosclerosis or aneurysmal dilatation of the involved vessel.1,7 Ingestion of aspirin and NSAIDs was common in this study (51%), similar to that described by Baettig et al.2 (43%). This is not surprising in this elderly group of patients, many with cardiovascular disease. Juler et al.7 have postulated NSAID-induced gastritis with subsequent vessel wall necrosis in the pathogenesis of the Dieulafoy lesion rupture. However, in the current study the VOLUME 50, NO. 6, 1999
I Norton, B Petersen, D Sorbi, et al.
adjacent mucosa in all cases was normal at endoscopic assessment. Initial endoscopy was diagnostic in 63% of cases. This is in part due to the fact that patients presenting with hematochezia routinely undergo upper endoscopy in the first instance to exclude a proximal lesion prior to lavage for colonoscopy. Thus, all colonic Dieulafoy lesions required at least 2 endoscopic procedures for diagnosis. Excluding the initial upper GI endoscopy in patients with colonic lesions, the first endoscopy was diagnostic in 73%. A previous study has reported a sensitivity of 49% for initial endoscopy.23 In view of the intermittent nature of bleeding from the Dieulafoy lesion, the sensitivity of endoscopic diagnosis is likely to be increased through early endoscopy in patients with acute GI bleeding. One third of Dieulafoy lesions in this study were extragastric. Although there are numerous published descriptions of lesions in nongastric sites,12-21 no other series has reported this high proportion of extragastric Dieulafoy lesions. We suspect that the traditional association of Dieulafoy lesions with the proximal stomach yields a reluctance to use the term for lesions located at extragastric sites. It is likely that these are often classified as ulcer disease in the duodenum or vascular malformation in the duodenum or other sites. Endoscopic management in this series was highly successful, with primary hemostasis achieved in 96%. This is identical to the rate of primary hemostasis reported by Baettig et al.2 The only endoscopic complication in this study was gastric perforation following injection of sodium tetradecyl sulfate. We no longer use sclerosant injection for treatment of these lesions, preferring epinephrine injection followed by either band ligation or thermal ablation using either heat probe or bipolar probe. Endoscopic ablative therapy has been advocated in a recent review of the Dieulafoy lesion, although no randomized controlled trials have been performed.1 Many reports have described successful hemostasis utilizing a variety of endoscopic modalities, such as injection with sclerosant2,3 or cyanoacrylate glue,24 thermal ablation with heat probe,4,25 neodymiumyttrium aluminum garnet laser,26 monopolar27,28 or bipolar probe,4,9 and band ligation.10,29 Most experience is with sclerosant injection and thermal ablation. Contact thermal ablation (e.g., heat probe) has been shown in this and other studies to be the mainstay of management. The prior use of epinephrine injection may be a useful adjunct to slow or stop bleeding prior to ablative therapy or prior to clot guillotine. We do not advocate the use of epinephrine alone in the management of suspected Dieulafoy GASTROINTESTINAL ENDOSCOPY
765
I Norton, B Petersen, D Sorbi, et al.
lesion.4,25,27,28 Use of neodymium-yttrium aluminum garnet laser in our study was early in the course of the study and is no longer used due to its cumbersome nature.23 Nonthermal treatment such as hemoclip may play a role in patients with bleeding diathesis, where minimization of tissue damage is desirable. No study has compared the results of surgical and endoscopic therapy for Dieulafoy lesion. A randomized study comparing endoscopic and surgical modalities would be impractical, given the relative infrequency of Dieulafoy lesions and the success of endoscopic therapies in uncontrolled series. This study demonstrates that when endoscopic therapy is unsuccessful the endoscopic examination still has an important role for identifying the lesion, because a nonbleeding Dieulafoy lesion may be undetectable at gastrotomy. This has led some authors to advocate a combined endoscopic and surgical approach to the patient requiring operation.30,31 Furthermore, if obliterative endoscopic therapy is not undertaken or is unsuccessful, it may be useful to tattoo the site of the bleeding lesion to guide the surgeon. Nine percent of subjects had further bleeding during the same hospital stay following endoscopic therapy. However, this was controlled at repeat endoscopy in all cases. This is comparable to the rates of recurrent bleeding in other large series including 11% in the series of Asaki et al.3 (46 patients), 11% in that of Baettig et al.2 (28 patients) and 22% in the series of Pointner et al.4 (11 patients). A review by Reilly and Al-Kawas23 of 177 Dieulafoy lesions quoted an overall rate of recurrent bleeding of 15%. Surgical intervention was required in 3 patients in whom bleeding could not be controlled endoscopically. This is consistent with three recent studies in which the rate of surgical intervention was between 2% and 6%.2-4 Angiography with Gelfoam embolization was reported to successfully arrest hemorrhage in 3 of 4 patients in one study.23 None of the patients in our cohort underwent angiographic therapy. It is likely that this modality may be particularly useful when endoscopic therapy has failed and the patient is deemed a poor candidate for surgery. Thirty-day mortality in this study was high (12 patients, 13%). Four early deaths were directly attributable to hemorrhage or its consequences. Eight additional deaths occurred within 30 days of the index bleeding episode but were not thought to be related to the hemorrhage or its consequences. During the follow-up period (median of 19 months), mortality was high (42%). However, no further death was directly related to the index episode 766
GASTROINTESTINAL ENDOSCOPY
Management and long-term prognosis of Dieulafoy lesion
or its sequelae. Similar rates were noted in another study, with 12% in-hospital and 45% overall mortality noted during a 14-month follow-up period.2 These investigators also believed that the mortality was largely a reflection of patient comorbidity. Due to the relative infrequency of this diagnosis, few large series of patients with extended follow-up have been reported. In this study, following definitive therapy and discharge from hospital, only 1 patient suffered a relapse of bleeding from a Dieulafoy lesion. This was one of the few patients who did not undergo “obliterative” therapy by either endoscopic or surgical means. The patient in question had an operative oversew of the original site and bled at the same site 6 years after the index episode. In summary, Dieulafoy lesion appears to account for approximately 2% of acute GI hemorrhage. Endoscopic therapy is effective and well tolerated and is of long-term benefit. No patients in our series have bled from more than 1 Dieulafoy lesion. Thirtyday and subsequent mortality following hemorrhage from a Dieulafoy lesion is high but relates almost exclusively to the patient’s comorbid status. ACKNOWLEDGMENTS The other members, both past and present, of our GI Bleed Team (1988-1999) are: Drs. David A. Ahlquist, Jeffrey A. Alexander, Todd H. Baron, Jonathan E. Clain, Rollin W. Hughes, Patrick S. Kamath, Mark V. Larson, Edward V. Lofthus, Darrell S. Pardi, John J. Poterucha, Kenneth W. Schroeder, William J. Tremaine, Thomas R. Viggiano, and Kenneth K. Wang. REFERENCES 1. Katz PO, Salas L. Less frequent causes of upper gastrointestinal bleeding [review]. Gastroenterol Clin North Am 1993;22:875-89. 2. Baettig B, Haecki W, Lammer F, Jost R. Dieulafoy’s disease: endoscopic treatment and follow up. Gut 1993;34:1418-21. 3. Asaki S, Sato H, Nishimura T, Ohkubo S, Yamagata R, Ito S, et al. Endoscopic diagnosis and treatment of Dieulafoy’s ulcer. Tohuko J Exp Med 1988;154:135-41. 4. Pointner R, Schwab G, Konigsrainer A, Dietze O. Endoscopic treatment of Dieulafoy’s disease. Gastroenterology 1988;94: 563-6. 5. Gallard T. Miliary aneurisms of the stomach giving cause to fatal hematemesis. Bull Soc Med Hop Paris 1894;1:84-91. 6. Dieulafoy G. Exulceration Simplex: L’intervention chirurgicale dans la hematemeses foudroyantes consecutives a l’exulceration simple de l’esomac. Bull Acad Med 1898;39:49-84. 7. Juler GL, Labitzke HG, Lamb R, Allen R. The pathogenesis of Dieulafoy’s gastric erosion. Am J Gastroenterol 1984;79:195200. 8. Molnar P, Miko T. Multiple arterial caliber-persistence resulting in hematomas and fatal rupture of the gastric wall. Am J Surg Pathol 1982;6:83-6. 9. Boron B, Mobarham S. Endoscopic treatment of Dieulafoy hemorrhage. J Clin Gastroenterol 1987;9:518-20. VOLUME 50, NO. 6, 1999
Management and long-term prognosis of Dieulafoy lesion
10. Brown GR, Harford WV, Jones WF. Endoscopic band ligation of an actively bleeding Dieulafoy lesion. Gastrointest Endosc 1994;40:501-3. 11. Parra-Blanco A, Takahashi H, Mendez Jerez PV, Kojima T, Aksoz K, Kirihara K, et al. Endoscopic management of Dieulafoy lesions of the stomach: a case study of 26 patients. Endoscopy 1997;29:834-9. 12. Bejanin H, Boivin C, Dehni N, Pipien I, Bloch F, Bruneval P, et al. Dieulafoy hemorrhagic duodenal ulcer: first case histologically confirmed [letter]. Gastroenterol Clin et Biol 1995; 19:227-8. 13. Barbier P, Luder P, Triller J, Ruchti C, Hassler Stafford A. Colonic hemorrhage from a solitary minute ulcer: report of three cases. Gastroenterology 1985;88:1065-8. 14. Dy NM, Gostout CJ, Balm RK. Bleeding from the endoscopically-identified Dieulafoy lesion of the proximal small intestine and colon. Am J Gastroenterol 1995;90:108-11. 15. Eguchi S, Maeda J, Taguchi H, Kanematsu T. Massive gastrointestinal bleeding from a Dieulafoy-like lesion of the rectum. J Clin Gastroenterol 1997;24:262-3. 16. Goins WA, Chatman DM, Kaviani MJ. Massive lower gastrointestinal bleeding due to “Dieulafoy’s vascular malformation” of the jejunum: case report. J Natl Med Assoc 1995;87: 766-70. 17. Goldenberg SP, DeLuca VA Jr, Marignani P. Endoscopic treatment of Dieulafoy’s lesion of the duodenum. Am J Gastroenterol 1990;85:452-2. 18. Paraf F. Dieulafoy disease of the gallbladder: first case? [letter]. Gastroenterol Clin et Biol 1996;20:712. 19. Wetscher G, Schwab G, Glaser K, Fend F, Bodner E, Pointner R. Dieulafoy lesion in a congenital double pylorus [letter]. Endoscopy 1994;26:374-5. 20. Scheider DM, Barthel JS, King PD, Beale GD. Dieulafoy-like
VOLUME 50, NO. 6, 1999
I Norton, B Petersen, D Sorbi, et al.
21.
22.
23. 24.
25.
26. 27. 28. 29.
30.
31.
lesion of the distal esophagus. Am J Gastroenterol 1994;89: 2080-1. Richards WO, Grove-Mahoney D, Williams LF. Hemorrhage from a Dieulafoy type ulcer of the colon: a new cause of lower gastrointestinal bleeding. Am Surg 1988;54:121-4. Gostout CJ, Wang KK, Ahlquist DL, Clain JE, Hughes RW, Larson MV, et al. Acute gastrointestinal bleeding: experience of a specialized management team. J Clin Gastroenterol 1992;14:260-7. Reilly HF, Al-Kawas FH. Dieulafoy’s lesion: diagnosis and management. Dig Dis Sci 1991;36:1702-7. D’Imperio D, Papadia C, Baroncini D, Piemontese A, Billi P. N-butyl-2-cyanoacrylate in the endoscopic treatment of Dieulafoy ulcer [abstract]. Endoscopy 1995;27:216. Lin HJ, Lee FY, Tsai YT, Lee SD, Lee CH, Kang WM. Therapeutic endoscopy for Dieulafoy’s disease. J Clin Gastroenterol 1989;11:507-10. Al Kawas FH, O’Keefe J. Nd:YAG laser treatment of a bleeding Dieulafoy lesion. Gastrointest Endosc 1987;33:38-9. Bakka A, Rosseland AR. Massive gastric bleeding from exulceratio simplex (Dieulafoy). Acta Chir Scand 1986;152:285-8. Hoffman J, Beck H, Jensen HE. Dieulafoy’s lesion. Surg Gynecol Obstet 1984;159:537-40. Murray KF, Jennings RW, Fox VL. Endoscopic band ligation of a Dieulafoy lesion in the small intestine of a child. Gastrointest Endosc 1996;44:336-9. Grisendi A, Lonardo A, Della Casa G, Frazzoni M, Pulvirenti M, Ferrari AM, et al. Combined endoscopic and surgical management of Dieulafoy vascular malformation. J Am Coll Surg 1994;179:182-6. Mixter CG, Sullivan CA. Control of proximal gastric bleeding: combined laparoscopic and endoscopic approach. J Laparoendosc Surg 1992;2:105-9.
GASTROINTESTINAL ENDOSCOPY
767