Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991), pp. 1702-1707
Dieulafoy' s Lesion Diagnosis and Management H A R O L D F. REILLY, III, MD, and FIRAS H. AL-KAWAS, MD
A review of 177 cases of upper gastrointestinal hemorrhage due to Dieulafoy's lesion is reported. Dieulafoy's lesion is frequently responsible for severe and recurrent upper gastrointestinal hemorrhage. The lesion was predominantly found in the proximal stomach. Repeat endoscopies were needed in 33% of the patients in order to make the correct diagnosis. When preoperative diagnosis and localization were made, surgery was an effective therapeutic modality. Therapeutic endoscopy was successful in achieving permanent hemostasis in 85% of the reported cases. Re-treatment was needed in an additional 10% and surgical therapy in 5% of the cases. Therapeutic endoscopy should be considered initially in all patients. Surgical intervention and angiography with embolization may be effective options if endoscopic therapy is unsuccessful. KEY WORDS: upper gastrointestinal hemorrhage; exulceratio simplex.
Dieulafoy's lesion is a rare but potentially lifethreatening source of upper gastrointestional bleeding. We present a review of 177 cases reported in the English literature to help further explore some of the characteristics of the lesion and evaluate the various diagnostic and treatment options available. Although specific data are not consistently available in all of the studies reported, we feel that there is enough information available to draw certain conclusions and to explore the trends concerning the diagnosis and management of the lesion. The lesion was first described by Gallard in 1884 (1) and was later characterized by the French surgeon Dieulafoy in 1896 (2) as the "exulceratio simplex" since the mucosal defect was small, and the underlying artery was large but histologically normal. Characteristically, the lesion is described as an abnormally large submucosal arterial vessel Manuscript received December 26, 1990; revised manuscript received June 17, 1991; accepted July 8, 1991. From the Division of Gastroenterology, Department of Medicine, Georgetown University Medical Center, Washington, D.C. 20007. Address for reprint requests: Dr. Firas H. AI-Kawas, Division of Gastroenterology, Room 2118, Georgetown University Hospital, 3800 Reservoir Road, NW, Washington, D.C. 20007.
1702
that protrudes through a solitary, minute mucosal defect (2-5 mm) most often located in the upper portion of the stomach (1-5). It ruptures spontaneously and bleeds massively for unclear reasons (3-5). Detection can be difficult due to its small size, unique location, and solitary nature. There is usually no significant visible tissue injury, which adds to the diagnostic difficulty. Its incidence as a source of upper gastrointestinal bleeding ranges from 0.3% to 6.7% (3-6). Microscopic examination of the lesion reveals no evidence of aneurysm formation, atherosclerosis, arteritis, or surrounding inflammation (7-9). Most histologic descriptions of similar lesions in the literature have been consistent with this characterization. The characteristic clinical presentation is recurrent (often massive) hematemesis at times associated with melena, hematochezia, and hypotension. Of the 177 cases described in the literature, 28% of patients presented with hematemesis alone, 51% presented with hematemesis accompanied by melena, and in 18%, melena occurred alone (Table 1) (3-31). When specified, transfusion requirements were often large. In fact, most patients required resuscitation with at least three or more units of Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
0163-2116/91/1200-1702506.50/09 1991PlenumPublishingCorporation
DIEULAFOY'S LESION TABLE 1. CLINICAL FEATURESOF REPORTED CASES OF GASTROINTESTINALHEMORRHAGEDUE TO DIEULAFOY'S LESION Age (years) Range Mean
16-93
58
Sex (%) M
Mode o f presentation (%)* F
101(68) 48(32)
H + ME
61(51)
H
ME,
HEZ
33(28) 22(18) 4(3)
*H." hematemesis, ME: melena, HEZ: hematochezia.
packed red blood cells (3, 4, 6, 7, 10, 13, 14, 17, 22, 25, 26, 28, 32, 36, 37). A male preponderance exists. In 149 cases in which patient sex was recorded, 101 (68%) of the patients were males and 48 (32%) were females (Table 1). There have been inferences in the literature regarding associations with the use of alcohol and NSAIDS; however, no consistent evidence exists to strongly support these claims. The bleeding site, when identified, was often within 6 cm of the gastroesophageal junction in the cardia or fundus of the stomach (9). In our review (Table 2), approximately 98% of the lesions located in the stomach tended to be in the upper portion of the stomach. Specifically, 67% were located high in the body of the stomach and 25% were located in the gastric fundus (Table 2) (1, 3, 4, 6, 8-10, 12-29, 34-36). Its frequent occurrence in the proximal stomach may be explained by the nature of the blood supply to that portion of the stomach. The presence of unusually large submucosal arterial branches in the corpus and fundus of the stomach is due to the fact that submucosal arteries along the lessei" curvature originate directly from the left get~tric artery outside the stomach. In the rest of the stomach, however, the submucosal vessels arise from a branching submucosal plexus system and are thus much smaller in caliber (9, 30, 38). These anatomical considerations may play a role in the initial development of the lesion and in its clinical presentation. In fact, when resected specimens were available and described in the literature, the lesions grossly consisted of a protuberant, unusuTABLE 2. LOCATIONOF DIEULAFOY'S LESION IN STOMACHIN REPORTED SERIES* Location HB
N(%)
LC
GC
39 (45)
5(6)
PW
Other AW
12 (14) 2(2)
NS
F
A
5(6)
22 (25)
2(2)
*HB: high body, LC: lesser curvature, GC: greater-curvature, PW: posterior wall, AW: anterior wall, NS: not stated, F: fundus, A" antrum. Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
ally tortuous and abnormally wide artery located just beneath the muscularis mucosa in the submu, cosa (3, 4, 7-9, 12-16, 18). Microscopic examination revealed a thick-walled vessel with a prominent muscularis media. On occasion, fibrin covered the arterial defect. No external surrounding inflammatory reaction, ulceration, or aneurysm formation Was evident. There is considerable debate concerning these lesions. However, the most accepted theory is that a persistent caliber vessel in the submucosa is exposed by a small mucosal erosion leading to massive bleeding (39). The exact cause for mucosal injury is not clear; however, Miko and Thomazy (39) suggested that the thin mucosa that is fixed to the pulsating artery becomes progressively weaker leading ultimately to the formation of an erosion overlying the vessel. Recently, reports have identified similar lesions in the duodenal bulb (31), the jejunum (32, 33), and in the right colon (37). The clinical presentation of patients with lesions in the duodenal bulb and proximal jejunum were similar to those that occurred in the stomach, while patients with lesions in the middle or distal jejunum and right colon presented with massive rectal bleeding (3133). Most of these lesions were removed via surgical resection. The gross and microscopic appearance was similar to those lesions described in the stomach. The etiology remains unclear. Detection and identification of the Dieulafoy's lesion as the source of bleeding can often be difficult, especially since most present with massive bleeding (4, 9, 27, 30). Prior to the development of endoscopy, most diagnoses were made during surgical exploration or at autopsy. Esophagogastroduodenoscopy ( E G D ) i s a useful tool for the diagnosis of upper gastrointestinal bleeding. It also has been a sensitive and accurate means of identifying the Dieulafoy's lesion (28-30). Endoscopically, the lesion appears as an isolated protruding vessel without associated ulcer (Figure 1). In the cases reviewed, upper endoscopy identified the Dieulafoy's lesion in approximately 82% of patients. Forty-nine percent of the lesions were identified during the initial endoscopic examination, while 33% required a second EGD to confidently identify the lesion as the bleeding source (Table 3) (3, 4, 6, 10, 11, 21, 22, 25-29, 3i, 34-36). The source of bleeding was not identified in 18% of the patients undergoing endoscopy. These lesions were later identified during surgical exploration. An alternative diagnostic tool currently available to assist in the diagnosis of the Dieulafoy's lesion is
1703
REILLY AND AL-KAWAS
Fig 1. Nonbleeding Dieulafoy's lesion in midbody of stomach (A)and Fundus (B). (C) Actively bleeding Dieulafoy's lesion in antrum (reference 29),
angiography. Angi0graphy has been useful in a limited number of cases when initial endoscopy did not confidently reveal the bleeding source. In our review, angiography was used in 14 patients as an adjunctive diagnostic tool. It was helpful in localization of the bleeding site in 11 patients (3, 23-26, 31, 32, 37). Seven of the patients had lesions in the stomach, while four were located in one o'f three separate sites (ie, duodenal bulb, jejunum, and ascending colon).
1704
In the preendoscopic era, the overall mortality from bleeding Dieulafoy's lesions was quite high. Surgical treatment has been advocated in the past as the treatment of choice due to the massive and often recurrent bleeding characteristic of the lesion (9-i 1). The surgical technique involved the use of a wide gastrotomy approach, which allowed for the best visualization~ Once identified, the vessel was either ligated, oversewn, or removed via a gastric wedge or "wide" resection (3, 4, 9, 10, 12-23, 26, Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
DIEULAFOY'S LESION TABLE 4. DIEULAFOY'S LESION OUTCOME OF SURGERYAFTER ENDOSCOPIC DIAGNOSIS* Outcome (%) Reference
N
Surgery
R
D
1,11,21,27,34 11,21,26,30,35 6,11,27 2,6,36 27 Total
6 5 3 5 4 23
SL WR SG OS/VP LE
6 (100) 5 (100) 1 (33) 4 (80) 4 (100) 20 (86)
0 0 2 (67) 1 (20) 0 3 (14)
*R: recovered, D: died, SL: suture ligati0n, WR: wedge resection, SG: subtotal gastrectomy, OS/VP: oversewn, vagotomy, pyloroplasty, LE: local excision.
Fig. 1 Continued.
27, 30). "Blind" resection or subtotal gastrectomy has not been recommended because lesions were often not included in the resected specimen and massive rebleeding often occurred in the patient postoperatively. A review of early surgical series, which describe surgically treated patients without the use of endoscopy, reveals that 12/36 (33%) patients died either postoperatively, perioperatively, or before surgery could be performed (1220). However, recent surgical data suggest the mortality rate has been reduced substantially following the use of preoperative endoscopy to confirm the diagnosis (Table 4). The evolution of endoscopy has led to the development of various endoscopic treatment techniques that have expanded the available therapeutic options (40). Theoretically, if permanent hemostasis could be achieved with endoscopic treatment, it
would obviate the need for emergency surgical intervention. Transfusion requirements, morbidity, and mortality associated wit h major upper gastrointestinal bleeding, it is hoped, also would be diminished. Endoscopic modalities useful in the treatment of bleeding upper gastrointesti0nal lesions include multipolar (bipolar) electrocoagu!ation, monopolar electrocoagulation, heater probe, injection sclerotherapy, and laser photocoagulation alone or in various combinations (40-48). Successful hemostasis with these techniques in the bleeding peptic ulcer has prompted their application to the management of the Dieulafoy's lesion. All of these methods have been applied with some success in the treatment of this lesion (1, 2, 4-29). In the cases reviewed, endoscopic therapy was utilized in 79 patients. Permanent hemostasis was accomplished in 85% of the patients following the first therapeutic endoscopy session. Twelve patients (15%) rebled. Repeat endoscopic therapy was successful in eight patients (10%). Surgical intervention was needed in four patients (5%). Heater probe alone (6) and injection therapy alone (5) or in Combination with TABLE 5. ENDOSCOPICTREATMENTOF DIEULAFOY'S LESION IN STOMACH
Author (rej)
N
Method*
Rebleed (%)
Hoffrnan (10) Bakka (27) Boron (28) A1-Kawas (29) Pointer (11)
2 1 3 1 18
M M B,T YAG B,HP,I
1 (50) 1 (!00) 1 (33) 0 4 (22)
TABLE 3. NUMBER OF ENDOSCOPICPROCEDURESREQUIREDTO DIAGNOSE DIEULAFOY'S LESION
Asaki (5)
46
I
5 (10.8)
Procedure
N (%)
LiD (6) Total
8 79
HP, I
Initial EGD Multiple EGDs Dx at surgery
32 (49) 22 (33) 12 (18)
Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
Outcome Surgery Retreated Retreated Retreated 2 Surgery 2 Surgery i t Retreated 4
0 12 (15)
*M: monopolar, B: bicap, HP: heater probe, I: injection~ YAG: Nd:YAG laser. tRebleeding occurred from overlooked ulcer requiring surgery.
1705
REILLY AND AL-KAWAS
bipolar electrocoagulation (I 1, 28) are equally successful. However, patient numbers in the studies are somewhat limited, and none are randomized. The best way to deliver endoscopic therapy is not established. In the study by Pointer et al (11) therapy was delivered around the lesion prior to treating the vessel. However, Lin et al (6) had equally successful results by targeting the vessel initially. Therefore both methods appear to be effective as long as the vessel is treated. In addition, even though data are lacking, it may be preferable to preinject epinephrine 1:10,000 around the lesion prior to the application of sclerotherapy or heat to reduce the potential for bleeding during therapy. When using a YAG laser, evaporation of the vessel should be avoided by applying short bursts of high energy starting around the lesion as suggested by AI-Kawas and O'Keefe (29). Hemostasis appears to be long-lasting since patients in the series reviewed were followed for an average of 61 months (2-120 months) without any evidence of rebleeding. However, follow-up in a larger number of patients may be needed to evaluate the long-term impact of therapeutic endoscopy. Angiography is another therapeutic option available for treatment of Dieulafoy's lesion. The technique utilizes gel-foam embolization. Although reported experience with the utilization o f this technique in treatment of this lesion is extremely limited, three of four patients in this review were treated and had successful results (24, 25). In conclusion, Dieulafoy's lesion has been identified more frequently in recent years as a source of major gastrointestional bleeding. This may be the result of heightened physician awareness of its existence, thereby leading to a more meticulous search to locate the lesion. The continued evolution of endoscopy and the development of effective endoscopic hemostatic methods offer viable alternatives in the management of Dieulafoy's lesion. Evidence suggests that since the advent of the endoscopic era in the management of bleeding secondary to the lesion, as both a diagnostic and/or therapeutic modality, overall mortality has diminished. We believe, therefore, that endoscopy should be considered initially in the evaluation to help establish the diagnosis and subsequently as the initial mode of therapy when an experienced endoscopist is available. Angiography and surgery certainly are strong available options, depending on the clinical situation and the outcome of therapeutic endoscopy.
1706
ACKNOWLEDGMENTS
The authors thank Drs. David Fleischer and Stanley Benjamin for reviewing the manuscript and Mrs. Norah Slevin for her secretarial help. REFERENCES I. Gallard MT: Aneurysmes miliaires de l'estomac. Soc Med Hop Bull Mim Soc Med, Hop 84, 1884 2. Dieulafoy G: L'exulceratio simple. In Manuel de Pathologie Interne. Paris, Masson, 1908, pp 178-305 3. Broberg A, Ihre T, Pyk E, Raaschou-Nielson T: Exuiceration simplex as a conceivable cause of massive gastric hemorrhage. Surg Gynecol Obstet 154:186-188, 1982 4. Strong RW: Dieulafoy's disease: A distinct clinical entity. Aust NZ J Surg 54:337-339, 1984 5. Asaki S, Sato H, Nishimura T, Ohkubo S, Yamagata R, Ito S, Saito Y, Miyazaki S, Yaginumi N: Endoscopic diagnosis and treatment of Dieulafoy's ulcer. Tohoku J Exp Med 154:135-141, 1988 6. Lin HJ, Lee FY, Tsai YT, Lee SD, Lee CH, Kang WM: Therapeutic endoscopy for Dieulafoy's disease. J Clin Gastroenterol 11(5):507-510, 1989 7. Juler GL~ Labitzke HG, Lamb R, Allen R: The pathogenesis of Dieulafoy's gastric erosion. Am J Gastroenterol 79:195200, 1984 8. Goldman RL: Submucosal arterial malformation ("aneurysm") of the stomach with fatal hemorrhage. Gastroenterology 46:589-594, 1964 9. Veldhuyzen van Zanten SJO, Bartlesman JFWM, Schipper MEI, Tytgat GNT: Recurrent massive hematemesis from Dieulafoy vascular malformations--a review of 101 cases. Gut 27:213-222, 1986 10. Hoffmann J, Beck H, Jensen HE: Dieulafoy's lesion. Surg Gynecol Obstet 159:537-540, 1984 11. Pointer R, Schwab G, Konigsrainer e, Dietze O: Endoscopic treatment of Dieulafoy's disease. Gastroenterology 94:563566, 1988 12. Frank W: Hematemesis associated with gastric arteriosclerosis. Gastroenterology 7:231-240, 1946 13. Mallory T: Case 32402. N Engl J Med 235:524-527, 1946 14. Donaldson GA, Hamlin E Jr: Massive hematemesis resulting from rupture of a gastric artery aneurysm. Report of three cases. N Engl J Med 243:369-373, 1950 15. Millard M: Fatal rupture of gastric aneurysm. Arch Pathol 59:363-371, 1955 16. Antonie T: Arteriosclerosis of the arteries of the stomach. Med J Aust 48i210-211, 1961 17. Chapman I, Lapi N: A rare course of gastric hemorrhage. Arch Intern Med 112:347-351, 1963 18. Markby CEP: Massive hemorrhage from superficial gastric erosions. Br J Surg 52(9):685-691, 1965 19. Bongiovi JJ, Duffy JL: Gastric hemangioma associated with upper gastrointestional bleeding. Arch Surg 95:93-98, 1967 20.Rossi NP, Green EW, Pike JD: Massive bleeding of the upper gastrointestional tract due to Dieulafoy's erosion. Arch Surg 97:797-800, 1968 21. Richter RM: Massive gastric hemorrhage from submucosal arterial malformation. Am J Gastroenterol 64:324-326, 1975 22. Palmer ED, Boyce HW Jr: Sclerotic submucosal gastric artery: A source of hemorrhage. Am Surg 30(2):83-87, 1964 Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
D I E U L A F O Y ' S LESION 23. Gough MH: Submucosal arterial malformation of the stomach as the probable cause of recurrent severe hematemesis in a 16-year-old girl. Br J Surg 64:522-524, 1977 24. Sherman L, Shenoy SS, Satchidanand SK, Neumann PR, Barrios GG, Peer RM: Arteriovenous malformation of the stomach. Am J Gastroenterol 72:160-164, 1979 25. Helliwell M, Irving JD: Hemorrhage from gastric artery aneurysms. Br Med J 282:460-461, 1981 26. MacMortensen NJ, Mountford RA, Davis JD, Jeans WD: Dieulafoy's disease: A distinctive arteriovenous malformation causing massive gastric hemorrhage. Br J Surg 70:7678, 1983 27. Bakka A, Rosseland AR: Massive gastric bleeding from exulceratio simplex (Dieulafoy). Acta Chir Scand 152:285288, 1986 28. Boron B, Mobarhan S: Endoscopic treatment of Dieulafoy hemorrhage. J Clin Gastroenterol 9(5):518-520, 1987 29. A1-Kawas FH, O'Keefe J: Nd:YAG laser treatment of a bleeding Dieulafoy's lesion. Gastrointest Endosc 33(1):3839, 1987 30. Louwerens JWK, Gratama S, Zwaan A, van der Schaar H: Dieulafoy's erosion in the stomach as a result ofintraabdominal vascular anomaly. Br J Surg 75:489-490, 1988 31. McClave SA, Goldschmid S, Cunningham JT, Boyd W Jr: Dieulafoy's cirsoid aneurysm of the duodenum. Dis Dis Sci 33:801-805, 1988 32. Matuchansky C, Babin P, Abadie JC, Payen J, Gasquet C, Barbier J: Jejunal bleeding from a solitary large submucosal artery. Gastroenterology 75:110-113, 1978 33. Boix J, Humbert P, Fernandez-Llamazares J, Planas R, Ojanguren I, Salva JA: Dieulafoy malformation. Dig Dis Sci 33(11): 1496-1497, 1988 34. DeVirgilio C, Dubrow TJ, Robertson JM, Wackym PA, Pelikin PCD, Cobb S, Williams RA: Dieulafoy's lesion associated with truncus arteriosus type IV: An unusual source of upper gastrointestinal hemorrhage. Am J Gastroenterol 83:865-867, 1988 35. Okada M, Iida M, Fuchigami T, Ohgushi H, Omae T, Kimura Y, Kido Hi Submucosal arterial malformation of the
Digestive Diseases and Sciences, VoL 36, No. 12 (December 1991)
36.
37.
38.
39.
40. 41.
42. 43.
44.
45.
46.
47.
48.
stomach diagnosed endoscopically. Gastrointest Endosc 29(1):30-31, 1983 Sarles HE Jr, Schenkein JP, Hecht RM, Sanowski RA, Miller P: Dieulafoy's ulcer: A rare source of massive gastric hemorrhage in an 11-year-old girl: Case report and literature review. Am J Gastroenterol 79:930-932, 1984 Barbier P, Luder P, Triller J, Ruchti CH, Hassler H, Stafford A: Colonic hemorrhage from a solitary minute ulcer. Report of three cases. Gastroenterology 88;1065-1068, 1985 Barlow TE, Bentley FH, Walder DN: Arteries, veins and arteriovenous anastomoses in the human stomach. Surg Gynecol Obstet 93:657-671, 1951 Miko TL, Thomazy VA: The caliber persistent artery of the stomach: A unifying approach to gastric aneurysm, Dieulafoy's lesion and submucosal arterial malformation. Hum Pathol 19:914-921, 1988 Fleischer DE: Endoscopic therapy of upper gastrointestinal bleeding in humans. Gastroenterology 90:217-234, 1986 Papp JP: Endoscopic electrocoagulation in the management of upper gastrointestinal tract bleeding. Surg Clin North Am 62:797-806, 1982 Steele RJC: Endoscopic haemostasis for non-variceal upper gastrointestinal hemorrhage. Br J Surg 76:219-225, 1989 Moore JP, Silvis SE, Vennes JA: Evolution of bipolar electrocoagulation in canine stomachs. Gastrointest Endosc 24:148-151, 1978 Laine L: Multipolar electrocoagulation in the treatment of active upper gastrointestinal tract hemorrhage. N Engl J Med 316:1613-1617, 1987 Laine L: Multipolar electrocoagulation for the treatment of ulcers with non-bleeding visible vessels: A prospective controlled trial. Gastroenterology 94:A246, 1988 Asaki S: Endoscopic hemostasis of gastrointestinal hemorrhage by local application of absolute ethanol: A clinical study. Tohoku J Exp Med 141:373-383, 1983 Sugawa C, Fujita Y, Ikeda T, Walt AJ: Endoscopic hemostasis by local injection of ninety-eight per cent dehydrated ethanol. Surg Gynecol Obstet 162:159-163, 1986 SoehendraN: Injection ofnon-variceal bleedingin the upper gastrointestinal tract. Endoscopy 17:129-132, 1985
1707
Dieulafoy' s Lesion Diagnosis and Management H A R O L D F. REILLY, III, MD, and FIRAS H. AL-KAWAS, MD
A review of 177 cases of upper gastrointestinal hemorrhage due to Dieulafoy's lesion is reported. Dieulafoy's lesion is frequently responsible for severe and recurrent upper gastrointestinal hemorrhage. The lesion was predominantly found in the proximal stomach. Repeat endoscopies were needed in 33% of the patients in order to make the correct diagnosis. When preoperative diagnosis and localization were made, surgery was an effective therapeutic modality. Therapeutic endoscopy was successful in achieving permanent hemostasis in 85% of the reported cases. Re-treatment was needed in an additional 10% and surgical therapy in 5% of the cases. Therapeutic endoscopy should be considered initially in all patients. Surgical intervention and angiography with embolization may be effective options if endoscopic therapy is unsuccessful. KEY WORDS: upper gastrointestinal hemorrhage; exulceratio simplex.
Dieulafoy's lesion is a rare but potentially lifethreatening source of upper gastrointestional bleeding. We present a review of 177 cases reported in the English literature to help further explore some of the characteristics of the lesion and evaluate the various diagnostic and treatment options available. Although specific data are not consistently available in all of the studies reported, we feel that there is enough information available to draw certain conclusions and to explore the trends concerning the diagnosis and management of the lesion. The lesion was first described by Gallard in 1884 (1) and was later characterized by the French surgeon Dieulafoy in 1896 (2) as the "exulceratio simplex" since the mucosal defect was small, and the underlying artery was large but histologically normal. Characteristically, the lesion is described as an abnormally large submucosal arterial vessel Manuscript received December 26, 1990; revised manuscript received June 17, 1991; accepted July 8, 1991. From the Division of Gastroenterology, Department of Medicine, Georgetown University Medical Center, Washington, D.C. 20007. Address for reprint requests: Dr. Firas H. AI-Kawas, Division of Gastroenterology, Room 2118, Georgetown University Hospital, 3800 Reservoir Road, NW, Washington, D.C. 20007.
1702
that protrudes through a solitary, minute mucosal defect (2-5 mm) most often located in the upper portion of the stomach (1-5). It ruptures spontaneously and bleeds massively for unclear reasons (3-5). Detection can be difficult due to its small size, unique location, and solitary nature. There is usually no significant visible tissue injury, which adds to the diagnostic difficulty. Its incidence as a source of upper gastrointestinal bleeding ranges from 0.3% to 6.7% (3-6). Microscopic examination of the lesion reveals no evidence of aneurysm formation, atherosclerosis, arteritis, or surrounding inflammation (7-9). Most histologic descriptions of similar lesions in the literature have been consistent with this characterization. The characteristic clinical presentation is recurrent (often massive) hematemesis at times associated with melena, hematochezia, and hypotension. Of the 177 cases described in the literature, 28% of patients presented with hematemesis alone, 51% presented with hematemesis accompanied by melena, and in 18%, melena occurred alone (Table 1) (3-31). When specified, transfusion requirements were often large. In fact, most patients required resuscitation with at least three or more units of Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
0163-2116/91/1200-1702506.50/09 1991PlenumPublishingCorporation
DIEULAFOY'S LESION TABLE 1. CLINICAL FEATURESOF REPORTED CASES OF GASTROINTESTINALHEMORRHAGEDUE TO DIEULAFOY'S LESION Age (years) Range Mean
16-93
58
Sex (%) M
Mode o f presentation (%)* F
101(68) 48(32)
H + ME
61(51)
H
ME,
HEZ
33(28) 22(18) 4(3)
*H." hematemesis, ME: melena, HEZ: hematochezia.
packed red blood cells (3, 4, 6, 7, 10, 13, 14, 17, 22, 25, 26, 28, 32, 36, 37). A male preponderance exists. In 149 cases in which patient sex was recorded, 101 (68%) of the patients were males and 48 (32%) were females (Table 1). There have been inferences in the literature regarding associations with the use of alcohol and NSAIDS; however, no consistent evidence exists to strongly support these claims. The bleeding site, when identified, was often within 6 cm of the gastroesophageal junction in the cardia or fundus of the stomach (9). In our review (Table 2), approximately 98% of the lesions located in the stomach tended to be in the upper portion of the stomach. Specifically, 67% were located high in the body of the stomach and 25% were located in the gastric fundus (Table 2) (1, 3, 4, 6, 8-10, 12-29, 34-36). Its frequent occurrence in the proximal stomach may be explained by the nature of the blood supply to that portion of the stomach. The presence of unusually large submucosal arterial branches in the corpus and fundus of the stomach is due to the fact that submucosal arteries along the lessei" curvature originate directly from the left get~tric artery outside the stomach. In the rest of the stomach, however, the submucosal vessels arise from a branching submucosal plexus system and are thus much smaller in caliber (9, 30, 38). These anatomical considerations may play a role in the initial development of the lesion and in its clinical presentation. In fact, when resected specimens were available and described in the literature, the lesions grossly consisted of a protuberant, unusuTABLE 2. LOCATIONOF DIEULAFOY'S LESION IN STOMACHIN REPORTED SERIES* Location HB
N(%)
LC
GC
39 (45)
5(6)
PW
Other AW
12 (14) 2(2)
NS
F
A
5(6)
22 (25)
2(2)
*HB: high body, LC: lesser curvature, GC: greater-curvature, PW: posterior wall, AW: anterior wall, NS: not stated, F: fundus, A" antrum. Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
ally tortuous and abnormally wide artery located just beneath the muscularis mucosa in the submu, cosa (3, 4, 7-9, 12-16, 18). Microscopic examination revealed a thick-walled vessel with a prominent muscularis media. On occasion, fibrin covered the arterial defect. No external surrounding inflammatory reaction, ulceration, or aneurysm formation Was evident. There is considerable debate concerning these lesions. However, the most accepted theory is that a persistent caliber vessel in the submucosa is exposed by a small mucosal erosion leading to massive bleeding (39). The exact cause for mucosal injury is not clear; however, Miko and Thomazy (39) suggested that the thin mucosa that is fixed to the pulsating artery becomes progressively weaker leading ultimately to the formation of an erosion overlying the vessel. Recently, reports have identified similar lesions in the duodenal bulb (31), the jejunum (32, 33), and in the right colon (37). The clinical presentation of patients with lesions in the duodenal bulb and proximal jejunum were similar to those that occurred in the stomach, while patients with lesions in the middle or distal jejunum and right colon presented with massive rectal bleeding (3133). Most of these lesions were removed via surgical resection. The gross and microscopic appearance was similar to those lesions described in the stomach. The etiology remains unclear. Detection and identification of the Dieulafoy's lesion as the source of bleeding can often be difficult, especially since most present with massive bleeding (4, 9, 27, 30). Prior to the development of endoscopy, most diagnoses were made during surgical exploration or at autopsy. Esophagogastroduodenoscopy ( E G D ) i s a useful tool for the diagnosis of upper gastrointestinal bleeding. It also has been a sensitive and accurate means of identifying the Dieulafoy's lesion (28-30). Endoscopically, the lesion appears as an isolated protruding vessel without associated ulcer (Figure 1). In the cases reviewed, upper endoscopy identified the Dieulafoy's lesion in approximately 82% of patients. Forty-nine percent of the lesions were identified during the initial endoscopic examination, while 33% required a second EGD to confidently identify the lesion as the bleeding source (Table 3) (3, 4, 6, 10, 11, 21, 22, 25-29, 3i, 34-36). The source of bleeding was not identified in 18% of the patients undergoing endoscopy. These lesions were later identified during surgical exploration. An alternative diagnostic tool currently available to assist in the diagnosis of the Dieulafoy's lesion is
1703
REILLY AND AL-KAWAS
Fig 1. Nonbleeding Dieulafoy's lesion in midbody of stomach (A)and Fundus (B). (C) Actively bleeding Dieulafoy's lesion in antrum (reference 29),
angiography. Angi0graphy has been useful in a limited number of cases when initial endoscopy did not confidently reveal the bleeding source. In our review, angiography was used in 14 patients as an adjunctive diagnostic tool. It was helpful in localization of the bleeding site in 11 patients (3, 23-26, 31, 32, 37). Seven of the patients had lesions in the stomach, while four were located in one o'f three separate sites (ie, duodenal bulb, jejunum, and ascending colon).
1704
In the preendoscopic era, the overall mortality from bleeding Dieulafoy's lesions was quite high. Surgical treatment has been advocated in the past as the treatment of choice due to the massive and often recurrent bleeding characteristic of the lesion (9-i 1). The surgical technique involved the use of a wide gastrotomy approach, which allowed for the best visualization~ Once identified, the vessel was either ligated, oversewn, or removed via a gastric wedge or "wide" resection (3, 4, 9, 10, 12-23, 26, Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
DIEULAFOY'S LESION TABLE 4. DIEULAFOY'S LESION OUTCOME OF SURGERYAFTER ENDOSCOPIC DIAGNOSIS* Outcome (%) Reference
N
Surgery
R
D
1,11,21,27,34 11,21,26,30,35 6,11,27 2,6,36 27 Total
6 5 3 5 4 23
SL WR SG OS/VP LE
6 (100) 5 (100) 1 (33) 4 (80) 4 (100) 20 (86)
0 0 2 (67) 1 (20) 0 3 (14)
*R: recovered, D: died, SL: suture ligati0n, WR: wedge resection, SG: subtotal gastrectomy, OS/VP: oversewn, vagotomy, pyloroplasty, LE: local excision.
Fig. 1 Continued.
27, 30). "Blind" resection or subtotal gastrectomy has not been recommended because lesions were often not included in the resected specimen and massive rebleeding often occurred in the patient postoperatively. A review of early surgical series, which describe surgically treated patients without the use of endoscopy, reveals that 12/36 (33%) patients died either postoperatively, perioperatively, or before surgery could be performed (1220). However, recent surgical data suggest the mortality rate has been reduced substantially following the use of preoperative endoscopy to confirm the diagnosis (Table 4). The evolution of endoscopy has led to the development of various endoscopic treatment techniques that have expanded the available therapeutic options (40). Theoretically, if permanent hemostasis could be achieved with endoscopic treatment, it
would obviate the need for emergency surgical intervention. Transfusion requirements, morbidity, and mortality associated wit h major upper gastrointestinal bleeding, it is hoped, also would be diminished. Endoscopic modalities useful in the treatment of bleeding upper gastrointesti0nal lesions include multipolar (bipolar) electrocoagu!ation, monopolar electrocoagulation, heater probe, injection sclerotherapy, and laser photocoagulation alone or in various combinations (40-48). Successful hemostasis with these techniques in the bleeding peptic ulcer has prompted their application to the management of the Dieulafoy's lesion. All of these methods have been applied with some success in the treatment of this lesion (1, 2, 4-29). In the cases reviewed, endoscopic therapy was utilized in 79 patients. Permanent hemostasis was accomplished in 85% of the patients following the first therapeutic endoscopy session. Twelve patients (15%) rebled. Repeat endoscopic therapy was successful in eight patients (10%). Surgical intervention was needed in four patients (5%). Heater probe alone (6) and injection therapy alone (5) or in Combination with TABLE 5. ENDOSCOPICTREATMENTOF DIEULAFOY'S LESION IN STOMACH
Author (rej)
N
Method*
Rebleed (%)
Hoffrnan (10) Bakka (27) Boron (28) A1-Kawas (29) Pointer (11)
2 1 3 1 18
M M B,T YAG B,HP,I
1 (50) 1 (!00) 1 (33) 0 4 (22)
TABLE 3. NUMBER OF ENDOSCOPICPROCEDURESREQUIREDTO DIAGNOSE DIEULAFOY'S LESION
Asaki (5)
46
I
5 (10.8)
Procedure
N (%)
LiD (6) Total
8 79
HP, I
Initial EGD Multiple EGDs Dx at surgery
32 (49) 22 (33) 12 (18)
Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
Outcome Surgery Retreated Retreated Retreated 2 Surgery 2 Surgery i t Retreated 4
0 12 (15)
*M: monopolar, B: bicap, HP: heater probe, I: injection~ YAG: Nd:YAG laser. tRebleeding occurred from overlooked ulcer requiring surgery.
1705
REILLY AND AL-KAWAS
bipolar electrocoagulation (I 1, 28) are equally successful. However, patient numbers in the studies are somewhat limited, and none are randomized. The best way to deliver endoscopic therapy is not established. In the study by Pointer et al (11) therapy was delivered around the lesion prior to treating the vessel. However, Lin et al (6) had equally successful results by targeting the vessel initially. Therefore both methods appear to be effective as long as the vessel is treated. In addition, even though data are lacking, it may be preferable to preinject epinephrine 1:10,000 around the lesion prior to the application of sclerotherapy or heat to reduce the potential for bleeding during therapy. When using a YAG laser, evaporation of the vessel should be avoided by applying short bursts of high energy starting around the lesion as suggested by AI-Kawas and O'Keefe (29). Hemostasis appears to be long-lasting since patients in the series reviewed were followed for an average of 61 months (2-120 months) without any evidence of rebleeding. However, follow-up in a larger number of patients may be needed to evaluate the long-term impact of therapeutic endoscopy. Angiography is another therapeutic option available for treatment of Dieulafoy's lesion. The technique utilizes gel-foam embolization. Although reported experience with the utilization o f this technique in treatment of this lesion is extremely limited, three of four patients in this review were treated and had successful results (24, 25). In conclusion, Dieulafoy's lesion has been identified more frequently in recent years as a source of major gastrointestional bleeding. This may be the result of heightened physician awareness of its existence, thereby leading to a more meticulous search to locate the lesion. The continued evolution of endoscopy and the development of effective endoscopic hemostatic methods offer viable alternatives in the management of Dieulafoy's lesion. Evidence suggests that since the advent of the endoscopic era in the management of bleeding secondary to the lesion, as both a diagnostic and/or therapeutic modality, overall mortality has diminished. We believe, therefore, that endoscopy should be considered initially in the evaluation to help establish the diagnosis and subsequently as the initial mode of therapy when an experienced endoscopist is available. Angiography and surgery certainly are strong available options, depending on the clinical situation and the outcome of therapeutic endoscopy.
1706
ACKNOWLEDGMENTS
The authors thank Drs. David Fleischer and Stanley Benjamin for reviewing the manuscript and Mrs. Norah Slevin for her secretarial help. REFERENCES I. Gallard MT: Aneurysmes miliaires de l'estomac. Soc Med Hop Bull Mim Soc Med, Hop 84, 1884 2. Dieulafoy G: L'exulceratio simple. In Manuel de Pathologie Interne. Paris, Masson, 1908, pp 178-305 3. Broberg A, Ihre T, Pyk E, Raaschou-Nielson T: Exuiceration simplex as a conceivable cause of massive gastric hemorrhage. Surg Gynecol Obstet 154:186-188, 1982 4. Strong RW: Dieulafoy's disease: A distinct clinical entity. Aust NZ J Surg 54:337-339, 1984 5. Asaki S, Sato H, Nishimura T, Ohkubo S, Yamagata R, Ito S, Saito Y, Miyazaki S, Yaginumi N: Endoscopic diagnosis and treatment of Dieulafoy's ulcer. Tohoku J Exp Med 154:135-141, 1988 6. Lin HJ, Lee FY, Tsai YT, Lee SD, Lee CH, Kang WM: Therapeutic endoscopy for Dieulafoy's disease. J Clin Gastroenterol 11(5):507-510, 1989 7. Juler GL~ Labitzke HG, Lamb R, Allen R: The pathogenesis of Dieulafoy's gastric erosion. Am J Gastroenterol 79:195200, 1984 8. Goldman RL: Submucosal arterial malformation ("aneurysm") of the stomach with fatal hemorrhage. Gastroenterology 46:589-594, 1964 9. Veldhuyzen van Zanten SJO, Bartlesman JFWM, Schipper MEI, Tytgat GNT: Recurrent massive hematemesis from Dieulafoy vascular malformations--a review of 101 cases. Gut 27:213-222, 1986 10. Hoffmann J, Beck H, Jensen HE: Dieulafoy's lesion. Surg Gynecol Obstet 159:537-540, 1984 11. Pointer R, Schwab G, Konigsrainer e, Dietze O: Endoscopic treatment of Dieulafoy's disease. Gastroenterology 94:563566, 1988 12. Frank W: Hematemesis associated with gastric arteriosclerosis. Gastroenterology 7:231-240, 1946 13. Mallory T: Case 32402. N Engl J Med 235:524-527, 1946 14. Donaldson GA, Hamlin E Jr: Massive hematemesis resulting from rupture of a gastric artery aneurysm. Report of three cases. N Engl J Med 243:369-373, 1950 15. Millard M: Fatal rupture of gastric aneurysm. Arch Pathol 59:363-371, 1955 16. Antonie T: Arteriosclerosis of the arteries of the stomach. Med J Aust 48i210-211, 1961 17. Chapman I, Lapi N: A rare course of gastric hemorrhage. Arch Intern Med 112:347-351, 1963 18. Markby CEP: Massive hemorrhage from superficial gastric erosions. Br J Surg 52(9):685-691, 1965 19. Bongiovi JJ, Duffy JL: Gastric hemangioma associated with upper gastrointestional bleeding. Arch Surg 95:93-98, 1967 20.Rossi NP, Green EW, Pike JD: Massive bleeding of the upper gastrointestional tract due to Dieulafoy's erosion. Arch Surg 97:797-800, 1968 21. Richter RM: Massive gastric hemorrhage from submucosal arterial malformation. Am J Gastroenterol 64:324-326, 1975 22. Palmer ED, Boyce HW Jr: Sclerotic submucosal gastric artery: A source of hemorrhage. Am Surg 30(2):83-87, 1964 Digestive Diseases and Sciences, Vol. 36, No. 12 (December 1991)
D I E U L A F O Y ' S LESION 23. Gough MH: Submucosal arterial malformation of the stomach as the probable cause of recurrent severe hematemesis in a 16-year-old girl. Br J Surg 64:522-524, 1977 24. Sherman L, Shenoy SS, Satchidanand SK, Neumann PR, Barrios GG, Peer RM: Arteriovenous malformation of the stomach. Am J Gastroenterol 72:160-164, 1979 25. Helliwell M, Irving JD: Hemorrhage from gastric artery aneurysms. Br Med J 282:460-461, 1981 26. MacMortensen NJ, Mountford RA, Davis JD, Jeans WD: Dieulafoy's disease: A distinctive arteriovenous malformation causing massive gastric hemorrhage. Br J Surg 70:7678, 1983 27. Bakka A, Rosseland AR: Massive gastric bleeding from exulceratio simplex (Dieulafoy). Acta Chir Scand 152:285288, 1986 28. Boron B, Mobarhan S: Endoscopic treatment of Dieulafoy hemorrhage. J Clin Gastroenterol 9(5):518-520, 1987 29. A1-Kawas FH, O'Keefe J: Nd:YAG laser treatment of a bleeding Dieulafoy's lesion. Gastrointest Endosc 33(1):3839, 1987 30. Louwerens JWK, Gratama S, Zwaan A, van der Schaar H: Dieulafoy's erosion in the stomach as a result ofintraabdominal vascular anomaly. Br J Surg 75:489-490, 1988 31. McClave SA, Goldschmid S, Cunningham JT, Boyd W Jr: Dieulafoy's cirsoid aneurysm of the duodenum. Dis Dis Sci 33:801-805, 1988 32. Matuchansky C, Babin P, Abadie JC, Payen J, Gasquet C, Barbier J: Jejunal bleeding from a solitary large submucosal artery. Gastroenterology 75:110-113, 1978 33. Boix J, Humbert P, Fernandez-Llamazares J, Planas R, Ojanguren I, Salva JA: Dieulafoy malformation. Dig Dis Sci 33(11): 1496-1497, 1988 34. DeVirgilio C, Dubrow TJ, Robertson JM, Wackym PA, Pelikin PCD, Cobb S, Williams RA: Dieulafoy's lesion associated with truncus arteriosus type IV: An unusual source of upper gastrointestinal hemorrhage. Am J Gastroenterol 83:865-867, 1988 35. Okada M, Iida M, Fuchigami T, Ohgushi H, Omae T, Kimura Y, Kido Hi Submucosal arterial malformation of the
Digestive Diseases and Sciences, VoL 36, No. 12 (December 1991)
36.
37.
38.
39.
40. 41.
42. 43.
44.
45.
46.
47.
48.
stomach diagnosed endoscopically. Gastrointest Endosc 29(1):30-31, 1983 Sarles HE Jr, Schenkein JP, Hecht RM, Sanowski RA, Miller P: Dieulafoy's ulcer: A rare source of massive gastric hemorrhage in an 11-year-old girl: Case report and literature review. Am J Gastroenterol 79:930-932, 1984 Barbier P, Luder P, Triller J, Ruchti CH, Hassler H, Stafford A: Colonic hemorrhage from a solitary minute ulcer. Report of three cases. Gastroenterology 88;1065-1068, 1985 Barlow TE, Bentley FH, Walder DN: Arteries, veins and arteriovenous anastomoses in the human stomach. Surg Gynecol Obstet 93:657-671, 1951 Miko TL, Thomazy VA: The caliber persistent artery of the stomach: A unifying approach to gastric aneurysm, Dieulafoy's lesion and submucosal arterial malformation. Hum Pathol 19:914-921, 1988 Fleischer DE: Endoscopic therapy of upper gastrointestinal bleeding in humans. Gastroenterology 90:217-234, 1986 Papp JP: Endoscopic electrocoagulation in the management of upper gastrointestinal tract bleeding. Surg Clin North Am 62:797-806, 1982 Steele RJC: Endoscopic haemostasis for non-variceal upper gastrointestinal hemorrhage. Br J Surg 76:219-225, 1989 Moore JP, Silvis SE, Vennes JA: Evolution of bipolar electrocoagulation in canine stomachs. Gastrointest Endosc 24:148-151, 1978 Laine L: Multipolar electrocoagulation in the treatment of active upper gastrointestinal tract hemorrhage. N Engl J Med 316:1613-1617, 1987 Laine L: Multipolar electrocoagulation for the treatment of ulcers with non-bleeding visible vessels: A prospective controlled trial. Gastroenterology 94:A246, 1988 Asaki S: Endoscopic hemostasis of gastrointestinal hemorrhage by local application of absolute ethanol: A clinical study. Tohoku J Exp Med 141:373-383, 1983 Sugawa C, Fujita Y, Ikeda T, Walt AJ: Endoscopic hemostasis by local injection of ninety-eight per cent dehydrated ethanol. Surg Gynecol Obstet 162:159-163, 1986 SoehendraN: Injection ofnon-variceal bleedingin the upper gastrointestinal tract. Endoscopy 17:129-132, 1985
1707
