2146

Gastrointestinal Bleeding in StrokeEelco F.M. Wijdicks, MD; Jimmy R. Fulgham, MD; Kenneth P. Batts, MD

Background and Purpose Patients-with ischemic or hemor-rhagic stroke are at risk for systemic complications. Thereasons why gastrointestinal bleeding occurs after stroke areunknown and have intuitively been attributed to stress ulcers.No study to date has addressed causes of gastrointestinalhemorrhage in stroke.

Methods Between 1976 and 1994, 17 patients identifiedfrom the Mayo Clinic medical record system as having gastro-intestinal hemorrhage and ischemic stroke (n=14) or intra-cerebral hemorrhage (n=3) were reviewed for presentation,associated causes, and outcome. Results of the endoscopicprocedures were compiled, and available gastric biopsies werereviewed.

Results In 17 patients with gastrointestinal bleeding afterstroke, sudden hematemesis, a decrease in hemoglobin level,or orthostatic hypotension was found as a presenting feature.

Systemic complications endanger patients withstroke; the most critical are aspiration pneumo-nia and sepsis. After acute brain injury, stress

ulcers may develop from vagal hyperactivity resulting inincreased gastric acid secretion or from mucosal isch-emia.1 Gastrointestinal bleeding may occur and maypotentially be devastating in patients with stroke treatedwith anticoagulation.

The clinical relevance of gastrointestinal bleeding inmajor strokes is uncertain. When stress ulcers bleed,they may contribute to mortality rates in patients withlarge strokes, but whether prophylactic therapy with itsenormous costs reduces time in the intensive-care unit,transfusion rates, or surgery is not known.

Scant data are available regarding gastrointestinalhemorrhage in head injury and subarachnoid hemor-rhage.1"4 No study to date has addressed causes ofgastrointestinal hemorrhage in stroke.

We report a review of gastrointestinal hemorrhage in17 patients with stroke. We undertook this survey todetermine the potential causes and to identify manage-ment strategies for prophylaxis.

Subjects and MethodsWe reviewed the medical records of all patients with isch-

emic or hemorrhagic stroke who had a gastrointestinal hem-orrhage as a complication during hospital stay in both Mayo

Received July 15, 1994; final revision received August 22, 1994;accepted August 22, 1994.

From the Department of Neurology (E.F.M.W., J.R.F.) andDepartment of Laboratory Medicine and Pathology (K.P.B.), andthe Neurology Critical Care and Stroke Service (E.F.M.W.,J.R.F.), Saint Marys Hospital, Mayo Clinic and Mayo Foundation,Rochester, Minn.

Correspondence to E.F.M. Wijdicks, MD, Department of Neu-rology, W8A, Mayo Clinic and Foundation, 200 First St, SW,Rochester, MN 55905.

O 1994 American Heart Association, Inc.

One patient presented with massive hematemesis, exsan-guination, and cardiac arrest. Endoscopic findings were avail-able in 14 patients and included gastroesophageal erosions,hemorrhagic gastritis, and gastric ulcer. In one patient, anadenocarcinoma of the gastric cardia was found. Putativepathogenetic agents were found in 16 of 17 patients andincluded a long history of nonsteroidal anti-inflammatorydrugs (n=6), acetylsalicylic acid (n=3), grossly prolongedanticoagulation (n=4), Helicobacter pylori (n=2), and cortico-steroids (n=l) .

Conclusions Gastrointestinal bleeding after stroke is rarelysevere and may not contribute significantly to mortality. Med-ication-induced gastrointestinal hemorrhage may be underap-preciated in this setting. (Stroke. 1994^5:2146-2148.)

Key Words • gastrointestinal hemorrhage • stress,psychological • stroke

Clinic-affiliated hospitals (Saint Marys Hospital and Method-ist Hospital) from 1976 to 1994. We excluded patients withaneurysmal subarachnoid hemorrhage, patients with chronicalcohol abuse and liver cirrhosis, and patients with acutecoagulopathies. A clinically significant gastrointestinal hemor-rhage was diagnosed on the basis of hematemesis, a nasogas-tric return containing gross blood or coffee ground-like mate-rial, or melena. Results of the endoscopic procedures werecompiled, and available gastric biopsies (K.P.B.) were re-viewed in detail. Potential risk factors for a peptic ulcer orgastritis were recorded.

ResultsWe found 17 patients with a documented gastrointes-

tinal hemorrhage. The mean age was 78 years (range, 63to 96 years). In the study period, 16 612 patients withischemic or hemorrhagic strokes were admitted to thehospital or developed in-hospital stroke; therefore, ourstudy patients represent 0.1% of the total population.Of the 17 patients with gastrointestinal hemorrhage, 14patients had a CT scan-documented ischemic stroke(lacuna [1 patient]; middle cerebral artery upper orlower division [7]; middle cerebral artery stem [3]; pos-terior cerebral artery territory [1]; and basilar arteryterritory [2]). An intracerebral hemorrhage was seen in3 patients (putamen [1] and subdural localization [2]).The Glasgow coma sum scores on admission were 14 in10 patients, 13 in 2 patients, and <8 in 5 patients.

Gastrointestinal hemorrhages presented in all pa-tients within 2 weeks after admission. Among these 17patients, 7 patients presented with hematemesis, 7patients with melena, and 3 patients with a suddendecrease in hemoglobin level of >4 g/dL. Orthostatichypotension (denned as a decrease of >20 mm Hg inthe systolic blood pressure when measured sitting up)was noted in 2 patients. In 1 patient, massive hemate-mesis was followed by overt shock and cardiac arrest. In

by guest on April 1, 2018

http://stroke.ahajournals.org/D

ownloaded from

Wijdicks et al GI Bleeding in Stroke 2147

AC (3)ASA(1) ..•

Helico- / Helicobacter(i) ...•''bacter(1) / „.•- ' '

'•..... ASA(1) / _.....---•

Upper circle shows the different endoscopic findings in 17patients with gastrointestinal bleeding after stroke. Possiblecauses of gastric lesions are depicted in the mirror image. ACindicates arrticoagulatton; ASA, aspirin; and NSAID, nonsteroldalanti-inflammatory drug.

none of the other patients was gastrointestinal bleedinglife threatening at any point during the clinical course.

The endoscopic findings that were available in 14patients are summarized in the Figure. Hemorrhagicgastritis, erosive gastritis, or gastric ulcers were found in8 patients. In 1 patient, an adenocarcinoma of thegastric cardia was found. In another patient, a hiatalhernia was associated with mechanical erosion of fundicmucosa at the diaphragmatic esophageal hiatus, a lesionsometimes referred to as a "Cameron lesion."5 Noendoscopic abnormalities were found in 4 patients.Endoscopy had not been performed in 3 patients withonly minor hematemesis or hemorrhagic nasogastricreturn.

Upper gastrointestinal biopsy specimens had beentaken in only 3 patients. Two of these showed Helico-bacter /y/on-associated gastritis; histological evidenceof ulceration was present in one of these. The third caserevealed near-normal mucosa with focal erosion thatwas felt to represent mechanical trauma to the gastricfundus at the diaphragmatic esophageal hiatus in apatient with hiatal herniation.

Thus, review of history and histology revealed possi-ble risk factors for upper gastrointestinal hemorrhage inall except 1 patient: long-term use of nonsteroidal anti-inflammatory drugs (NSAID) (dose, £75 mg/d) (6 of17), grossly prolonged anticoagulation (4 of 17), aspirin(dose, 325 to 975 mg/d) (3 of 17), corticosteroids (1 of17), and biopsy-proven H pylori (2 of 17).

None of the patients required mechanical ventilation.No patients were treated with prophylactic antacids orhistamine H2-receptor antagonists. After gastrointesti-nal bleeding, 7 patients required a blood transfusion, allpatients were treated with cimetidine, and repeat hem-orrhage did not occur. Emergency surgery was notneeded.

DiscussionThe pathogenesis of upper gastrointestinal bleeding

after stroke is unclear; intuitively it has been connectedwith "stress." Erosive or hemorrhagic gastritis was afrequent endoscopic finding in our patients with gastro-intestinal bleeding after stroke. However, the cause ofthese changes may not be simply "stress," since alter-native explanations for the hemorrhage were present in

nearly all cases. Our findings suggest a strong associa-tion with previous long-term NSAID and aspirin use;however, without a control group a causal relation is notestablished. The significance of H pylori in 2 of 3patients is not clear, since Hpylori would be expected tobe relatively common in the elderly population.

One rationale for analyzing gastrointestinal hemor-rhage after stroke is to determine whether high-riskpatients can be identified and whether potentially costlyprophylactic treatment is indicated. Previous studiesconducted in neurosurgical populations or critically illpatients found that no specific risk factor could beidentified, although there was an increased incidence inpatients on mechanical ventilators and, not unexpect-edly, in patients with a coexisting coagulopathy.167 Ourpatient population, therefore, may not represent the fullspectrum of gastrointestinal bleeding with stroke be-cause the majority of patients were not mechanicallyventilated.

Stress-related gastric lesions generally occur in mul-tiple sites. They may be confined to the mucosa (ero-sions) or extend into submucosa or beyond (ulcers). Thepathophysiology of stress-related damage is not wellunderstood. Experimental models have shown thatstress activates the hypothalamus, resulting in choliner-gic stimulation to the stomach. Substances such asacetylcholine, histamine, and endogenous thyrotropin-releasing hormone may also increase the vulnerabilityof the mucosa.8 In addition, it has been stated thatstress-related lesions in the gastric mucosa cannot bedistinguished from medication-induced lesions.8 Sinceat least one mechanism by which NSAIDs cause gastricerosion/ulceration is by inhibition of prostaglandins, asimilar mechanism in stress is possible.

We documented a potential trigger other than stressfor upper gastrointestinal hemorrhage in the majority ofour patients. This suggests that stress from acute braininjury may in fact be quite infrequent or perhaps mayact as an additive risk factor in patients with otherpredispositions for upper gastrointestinal hemorrhage.The endoscopic and histological similarities betweenstress gastritis and NSAID-associated gastritis9 make itdifficult to distinguish these two processes. We did notrecognize any cases of endoscopically diagnosed erosiveor hemorrhagic gastritis in the absence of prior NSAIDuse. The presence of NSAID use in a large proportionof our patients is of interest,10-11 since they are well-established causes of gastric ulcers and erosions. Arecent cross-sectional study found erosive gastritis inabout 50% of patients who used NSAIDS.8 The signif-icance of Hpylori in 2 of 3 sampled patients is not clearbecause so few patients underwent biopsy. H pyloriwould be expected to be relatively common in theelderly, thus stroke-prone, population. The associationbetween Helicobacter jejuni and gastric ulcers has beendemonstrated in many studies, and risk of infectionincreases with age.9

In the vast majority of our patients gastrointestinalhemorrhage did not produce a life-threatening situa-tion. Mild hematemesis, melena, and sudden decreasein hemoglobin were important presenting signs, andtreatment with H2-receptor antagonist and a singleblood transfusion resolved this complication. Fortu-nately, massive bleeding was uncommon and was notobserved in patients with ischemic stroke on heparin

by guest on April 1, 2018

http://stroke.ahajournals.org/D

ownloaded from

2148 Stroke Vol 25, No 11 November 1994

during hospitalization. In patients who presented withclinical signs of gastrointestinal bleeding, upper endos-copy was often helpful because in most patients local-ization of the hemorrhage to the stomach could berealized by the finding of fresh blood with a demonstra-ble lesion. An exposed vessel on the posterior wall ofthe duodenum or on the lesser curve of the stomach,typically associated with high risk of rebleeding, wasabsent and reinforces conservative management.

It should be pointed out that in this study, patientswith very minor gastrointestinal hemorrhages may havebeen missed. Most patients in our series had fairlysignificant hemorrhages that resulted in endoscopicinvestigations. The true frequency of gastrointestinalhemorrhage in stroke therefore remains unknown.

However, the very low incidence of significant gastro-intestinal bleeding requiring blood transfusion in ourretrospective series precludes indiscriminate use of pro-phylactic therapy. In stroke, prophylactic treatment withagents that reduce gastric acidity, known to effectivelyreduce gastric bleeding, may be tailored toward patientswith a history of peptic ulcer disease or use of aspirin,NSAIDS, or corticosteroids.

References1. Chan KH, Mann KS, Lai ECS, Ngan J, Tuen H, Yue CP. Factors

influencing the development of gastrointestinal complications afterneurosurgery: results of multivariate analysis. Neurosurgery. 1989;25:378-382.

2. Fleischer DE. Endoscopic control of upper gastrointestinal bleeding.J Clin Gastroenterol. 1990;2:S41-S47.

3. Halloran LG, Zfass AM, Gayle WE, Wheeler CB, Miller JD.Prevention of acute gastrointestinal complications after severehead injury: a controlled trial of cimetidine prophylaxis. Am J Surg.1980;139:44-48.

4. Tanaka S, Mori T, Ohara H. Gastrointestinal bleeding in cases ofruptured cerebral aneurysms. Ada Neurochir (Wien). 1979;48:223-230.

5. Cameron AJ, Higgins JA. Linear gastric erosion: a lesion asso-ciated with large diaphragmatic hernia and chronic blood lossanemia. Gastroentewlogy. 1986;91:338-342.

6. Kamada T, Fusamoto H, Kawano S, Noguchi M, Hiramatsu K,Masuzawa M, Abe H, Fujii C, Sugimoto T. Gastrointestinalbleeding following head injury: a clinical study of 433 cases.J Trauma. 1977;17:44-47.

7. Cook DJ, Fuller HD, Guyatt GH, Marshall JC, Leasa D, Hall R,Winton TL, Rutledge F, Todd TJR, Roy P, LaCroix J, Griffith MS,Willan A, for the Canadian Critical Care Trials Group. Riskfactors for gastrointestinal bleeding in critically ill patients. N EnglJ Med. 1994;330:377-381.

8. Bresalier RS. The clinical significance and pathophysiology ofstress-related gastric mucosal hemorrhage. J Clin Gastroenterol.1991;2:S35-S43.

9. Laine L, Weinstein WM. Subepithelial hemorrhages and erosionsof human stomach. Dig Dis Sci. 1988;33:490-503.

10. Kurata JH. An assessment of nonsteroidal anti-inflammatory drugsas a risk factor in ulcer disease. Ann Intern Med. 1991;114:311-316.

11. Rauus EAJ, Langenberg W, Houthoff HT, Zanen HC, TytgatGNJ. Campylobacter pyloritis-associated chronic active antral gas-tritis: a prospective study of its prevalence and the effects ofantibacterial and antiulcer treatment. Gastroentewlogy. 1988;94:3-37.

12. Steele RJC, Chung SCS, Leung JWC. Practical Management ofAcute Gastrointestinal Bleeding. Stoneham, Mass: Butterworth-Heinemann; 1993.

by guest on April 1, 2018

http://stroke.ahajournals.org/D

ownloaded from

E F Wijdicks, J R Fulgham and K P BattsGastrointestinal bleeding in stroke.

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 1994 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.25.11.2146

1994;25:2146-2148Stroke. 

http://stroke.ahajournals.org/content/25/11/2146World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer available in the

Permissions in the middle column of the Web page under Services. Further information about this process isOnce the online version of the published article for which permission is being requested is located, click Request

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Stroke Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on April 1, 2018

http://stroke.ahajournals.org/D

ownloaded from