CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2003;1:51–56

Sucralfate Therapy in NSAID Bleeding Gastropathy

JUAN–R. MALAGELADA,* ARTURO RODRÍGUEZ DE LA SERNA,‡ HANS GERT DAMMANN,§

MIQUEL PONS,㛳 CARLOS ARMAS,¶ MIQUEL SALA,# XAVIER TENA,** ENRIQUE CELDRÁN,‡‡
and ANTONI MESA‡‡
*Department of Gastroenterology, Hospital Vall d’Hebron, Autonomous University of Barcelona, Barcelona; ‡Department of Internal Medicine,
Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; §Wissenschaftliches Institut, Hamburg, Germany; 㛳Department of Rheumatology,
Hospital Municipal de Badalona, Badalona; ¶Department of Rheumatology, Hospital Clı́nico Universitario “Lozano Blesa,” Zaragoza;
#Department of Internal Medicine, Hospital de Figueres, Figueres; **Department of Rheumatology, “Germans Trias i Pujol,” Badalona; and
‡‡Department of Clinical Research, Merck Farma y Quı́mica, S.A., Mollet del Vallés, Barcelona, Spain

Background & Aims: A randomized, double-blind, place- tions.4 – 8 Inhibition of gastroduodenal prostaglandin
bo-controlled, multicenter study was conducted to as- (PG) synthesis by NSAIDs9 impairs PG-dependent mu-
sess the efficacy of 2 g sucralfate suspension in treating cosal protective mechanisms and facilitates the develop-
gastric mucosal lesions caused by long-term treatment ment of ulcerative lesions with a high rate of complica-
with nonsteroidal anti-inflammatory drugs (NSAIDs). tions, such as hemorrhage and perforation.3,10,11
Methods: Only patients given NSAIDs continuously for at
Dyspepsia or gastrointestinal intolerance is the most
least 2 months with positive fecal occult blood (FOB)
and endoscopically confirmed mild to moderate muco-
common cause of NSAID withdrawal,12 and about 10%–
sal lesions (Lanza scale, grades 2– 4) were included. 12% of all patients given NSAIDs discontinue treatment
After 1-week run-in phase, patients were stratified into 2 because of direct adverse effects including mucosal dam-
groups according to gastropathy-related symptoms dur- age and bleeding.13,14 The prevalence of severe gastroin-
ing the preceding 7 days (symptomatic vs. asymptom- testinal lesions caused by ingestion of NSAIDs varies
atic) and randomized to 2 g (10 mL) of sucralfate sus- between 10% and 30% for the gastric mucosa and be-
pension or placebo twice a day over a 6-week period. tween 2% and 19% for the duodenal mucosa,15 and the
NSAIDs were given according to each patient’s dosage risk is dose-dependent. The development of new cyclo-
schedule and always after meals. Results: Twenty-five oxygenase (COX) 2–specific inhibitors offers promise of
patients received sucralfate and 25 received placebo. At
NSAIDs with improved gastrointestinal tolerability, but
the end of the study, 68% (17/25) of patients given
the use of these new compounds has not become gener-
sucralfate had no lesions (Lanza grade 0) on endoscopy
compared with 35% (8/23) in controls (P ⴝ 0.042). The
alized yet.
Lanza grades in patients given sucralfate were signifi- Sucralfate, the aluminum salt of sucrose octosulfate, is
cantly improved compared with the placebo patients an antiulcer drug that increases the release of endogenous
(P ⴝ 0.022). Conclusions: In this target population se- PGE2 from the gastric mucosa16,17 and exerts a trophic
lected according to positive FOB test and endoscopic effect by activating genes encoding for epidermal growth
evidence of mucosal injury, chronic administration of factor and its receptors.18,19 Thus, its mode of action is
sucralfate significantly decreased NSAID-induced gastric different from alkalinizing agents, such as antacids or
erosions. antisecretory agents. It can be described as mucosal-
protective because it strengthens the natural defense
onsteroidal anti-inflammatory drugs (NSAIDs) are mechanisms of the gastrointestinal tract and because it
N widely used to treat pain and inflammation, with
more than 700 million prescriptions written worldwide
seems to protect the ulcerated area against attack by acid
and pepsin.
every year.1 The therapeutic benefits of NSAIDs are well It has been shown that sucralfate reduces mucosal
established, and, as a rule, these agents possess an en- damage from various strong irritants including etha-
hanced safety margin. However, widespread use in in- nol20,21 and NSAIDs.16,22,23 Accordingly, it may be pos-
dustrialized countries has led these drugs to become the
leading cause of serious adverse events following a phar- Abbreviations used in this paper: FOB, fecal occult blood; NSAID,
macologic therapy.2,3 Gastroduodenal toxicity accounts nonsteroidal anti-inflammatory drug; PG, prostaglandin.
© 2003 by the American Gastroenterological Association
for 21%–25% of reported adverse reactions in patients 1542-3565/03/$35.00
taking these drugs in combination with other medica- doi:10.1053/jcgh.2003.50008
52 MALAGELADA ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 1, No. 1

tulated that this agent may be useful to prevent gastric treatment with H2-receptor antagonists, proton pump inhib-
mucosal lesions in patients taking NSAIDs. In fact, the itors, and cytoprotective drugs; history of gastric surgery;
effectiveness of sucralfate coadministered with NSAIDs history of portal hypertension (hepatic encephalopathy, esoph-
in long-term therapeutic regimens has been assessed in a ageal varices, and/or portocaval shunt); and drug or alcohol
few placebo-controlled studies.24 –26 In these studies, abuse in the past 5 years or concurrent consumption of more
however, the presence of peptic and nonpeptic symptoms than 60 mg ethanol per day. Ingestion of low doses of antacid
as needed for symptomatic relief was allowed (see later). All
at baseline was an indispensable criterion for patients’
patients provided written informed consent and the study
enrollment and elective endoscopic assessment. Unfortu-
protocol was approved in advance by the Institutional Review
nately, it also resulted in heterogeneous patient groups, Boards or ethical committees of the participating centers. The
many of whom did not have significant gastroduodenal trial was conducted in accordance with the Declaration of
lesions, only symptoms. To prevent selection bias related Helsinki (Hong Kong amendments) and the standards of
to the presence of upper gastrointestinal symptoms, we “Good Clinical Practice for Trials on Medicinal Products in the
designed the present study to include only patients with European Community” (approved by the Committee for Pro-
positive fecal occult blood and endoscopically confirmed prietary Medical Products on July 11, 1990).
mild to moderate mucosal lesions. Thus, we have evalu-
ated the efficacy of sucralfate in the healing of gastric Study Procedures
bleeding lesions induced by chronic treatment with Eligible patients underwent examination of stool for
NSAIDs in both symptomatic and asymptomatic indi- occult blood. Those patients who had a positive fecal occult
viduals, all of whom had proven mucosal injury. blood test underwent endoscopic assessment of gastroduodenal
lesions, evaluation of gastropathy-related symptoms, physical
Materials and Methods examination, and laboratory tests during a 1-week run-in
period. Fecal occult blood (FOB) test was performed on sam-
A randomized, double-blind, placebo-controlled, mul-
ples of 3 different stools using a commercial diagnostic kit
ticenter, multinational, phase III study was conducted to
evaluate the efficacy of 2 g (10 mL) sucralfate suspension, (HemoFEC; Boehringer Mannheim, Mannheim, Germany).
administered orally twice a day, in treating gastric mucosal Bleeding was considered to be actually present when at least 2
lesions caused by long-term treatment with NSAIDs. of the 3 samples were positive. Endoscopy was performed after
an overnight fast with either pharyngeal anesthesia or intra-
Subject Population venous sedation. Erosion was defined as a mucosal break usu-
Between October 1996 and June 1998, patients with a ally of ⬍5 mm in diameter without definite depth in the
definitive diagnosis of rheumatoid arthritis, ankylopoietic stomach, duodenum, or both. Erosions were assessed with the
spondylitis, osteoarthritis, or psoriatic arthritis attending the modified Lanza scale12 and defined as grade 0, no lesion; grade
outpatient clinics of the Services of Rheumatology of 7 acute- 1, presence of 1 lesion; grade 2, 2– 4 lesions; grade 3, 5–10
care teaching hospitals or practices in Spain and Germany were lesions; grade 4, ⬎10 lesions; and grade 5, ulcer (mucosal
eligible to participate in the study provided that oral or rectal breaks of at least 5 mm in diameter with unequivocal depth in
NSAIDs within a predetermined dose range had been contin- the stomach, duodenum, or both). The participating endosco-
uously taken for at least 2 months before the study. The pists at each center were experienced and the Lanza scoring
allowed daily doses of the commonly used NSAIDs were as system discussed among them, but no specific training mate-
follows: 100 –150 mg of diclofenac, 500 –1000 mg of rial was used.
naproxen, 75–150 mg of indomethacin, 600 –1800 mg of The investigator asked patients whether they had heartburn,
ibuprofen, 10 –20 mg of piroxicam, 60 –120 mg of acemeta- abdominal fullness, nausea/vomiting, epigastric pain/cramp-
cin, 10 –20 mg of tenoxicam, 300 – 600 mg of etodolac, 100 – ing, and loss of appetite during the preceding 7 days. Mild or
200 mg of ketoprofen, and 7.5–15 mg of meloxicam. occasionally mild gastropathy-related symptoms were graded
Patients of both sexes and of any ethnic group, above 40 as 1, moderate or occasionally moderate as 2, often moderate as
years of age (postmenopausal or surgically sterilized women or 3, occasionally severe as 4, and often severe as 5. Patients who
women using a reliable contraceptive method), in whom there scored ⱖ5 points (sum of the scores of the 5 symptoms) were
were no anticipated changes in their anti-inflammatory ther- considered symptomatic, whereas patients with ⱕ4 points
apy schedule within the next 10 weeks were included. Exclu- were regarded as asymptomatic. Those patients scoring 20
sion criteria were as follows: known hypersensitivity to the points or higher were not randomized and withdrawn. Other
study medication including NSAIDs or history of drug-in- investigations included antibody testing for Helicobacter pylori
duced anaphylactic shock; concurrent treatment with steroids status (ELISA); routine laboratory tests; barium enema exam-
at a daily dose ⱖ10 mg of prednisone (or its equivalent); severe ination or colonoscopy; and measurements of body weight,
gastrointestinal disorders including carcinoma, Zollinger-Elli- pulse rate, and blood pressure using a standard cuff sphygmo-
son syndrome, active peptic ulcer, and bleeding esophagitis; manometer. Diastolic blood pressure was determined as Ko-
hematemesis, melena, or gastrointestinal perforation; antiulcer rotkoff sound V.
January 2003 CYTOPROTECTION OF SUCRALFATE 53

Patients with negative FOB test and/or endoscopically con- gastrointestinal bleeding, or perforation; assessment of erosions
firmed gastric carcinoma or Lanza grade 0, 1, or 5 lesions were or hemorrhagic lesions in the duodenum by endoscopy at the
excluded as were patients with malignant tumors of the colon end of the study; amount of antacids; and occurrence of adverse
and rectum, renal impairment (serum creatinine concentration events.
1.5 times above the upper normal limit), liver dysfunction According to an estimated healing rate among placebo-
(twice or higher above the upper normal limit of at least 2 of treated patients of 40%, a total of 120 patients (60 per
the following parameters: plasma albumin level, serum alanine treatment group) would be required to detect differences as
and aspartate aminotransferases, alkaline phosphatase, and small as 30% in the healing rate among treatment groups with
␥-glutamyl transpeptidase); uncontrolled hypertension with a statistical power of 1-␤ ⫽ 0.9 and ␣ level of 0.05. The
diastolic blood pressure ⬎100 mm Hg; and those requiring primary efficacy analysis used an intention-to-treat approach
iron administration with hemoglobin values ⬍10 g/dL. that included all patients meeting inclusion criteria who took
at least one dose of medication. The Fisher exact test, the exact
Randomization and Treatment Phase test for 2 ⫻ c contingency tables, and the Wilcoxon rank-sum
After the 1-week run-in phase (screening visit, days test were used for statistical analysis. Data were analyzed with
⫺7 to ⫺1), patients were stratified into 2 groups according to the Statistical Analysis Systems (SAS) software package and
gastropathy-related symptoms (symptomatic vs. asymptom- with StatXact (Cytel). The level of significance was set at P ⬍
atic) and randomized to treatment with sucralfate or placebo 0.05 based on two-tailed testing.
by a computer-generated randomization list. Investigators at
each center contacted by phone the coordinating center for
Results
allocation of treatment number. Telephone randomization en-
sured that the final groups were balanced with regard to Configuration of Study Population
history of peptic ulcer. Patients were instructed to take the The study was stopped when 51 patients had been
study medication, i.e., 5 mL sucralfate suspension (1 g sucral-
recruited because entry rate was slow. However, the
fate) or 5 mL placebo suspension with the same organoleptic
blinding of the study was maintained until the statistical
characteristics, twice a day, the first dose (2 sachets) before
breakfast and the second dose (2 sachets) before dinner. analysis was performed as described in the study proto-
NSAIDs were given according to each patient’s dosage sched- col. One eligible patient scored Lanza grade 1 for the
ule and always after meals. The NSAIDs were continued gastric mucosal lesion and Lanza grade 0 for the duode-
without dose reduction through the entire study period. Al- num on endoscopic examination at the screening visit
lowed concomitant medications included intramuscular gold and was not randomized.
salts, paracetamol, and antacids (magaldrate 400 mg) at the Thus, the study population consisted of 50 patients,
physician’s discretion. The duration of the study was 6 weeks 25 assigned to the group of treatment with sucralfate and
from the time of initial drug administration (visit 2). 25 to placebo. The main target variable was missing for
Assessments were carried out at 2-week intervals with in- 2 patients randomized to placebo. Treatment groups
terim visits every 7 days. At each visit, patients received a box were comparable with regard to baseline characteristics.
containing the randomized study medication for the next
History of peptic ulcer was recorded in 6 (24%) patients
14-day treatment period including 28 extra sachets as a reserve
in the sucralfate group and in 4 (16%) in the placebo
medication. Compliance was assessed by counting the number
of sachets the patient returned. Gastropathy-related symptoms group. On the other hand, gastropathy-related symptoms
were assessed at each visit, whereas FOB test and endoscopy were present in 14 (56%) assigned to sucralfate treat-
were repeated at the end of the study. The presence of adverse ment and in 13 (52%) allocated to placebo. Antibody
events as well as their severity and putative relation with the testing for Helicobacter pylori status was positive in 20
study medication were assessed by the investigator at each cases (sucralfate 9, placebo 11) and negative in the re-
follow-up assessment. maining 30 (sucralfate 16, placebo 14).
Statistical Analysis Treatment Outcome
The primary efficacy end point was to compare Lanza Results of treatment are shown in Table 1. At the
grades of the gastric mucosa assessed by endoscopy at the end end of the study, 42 days after randomization, lesions in
of the study between sucralfate- and placebo-treated patients.
the stomach at endoscopy as measured by the Lanza scale
Secondary efficacy end points included a comparison of the
following variables between patients in both treatment groups:
were significantly reduced in patients given sucralfate
healing success defined as Lanza grade 0 at the end of the compared with patients given placebo (P ⫽ 0.022).
study; rate of FOB negative test at the end of the study; mean Accordingly, the healing rate for patients with grade 0
scores of each gastropathy-related symptom and the sum of was significantly lower among placebo-treated patients
symptom scores for the subset of symptomatic patients; inci- than among patients given sucralfate (35 vs. 68%, P ⫽
dence of gastric complications, such as ulcer, upper or lower 0.042). Changes in gastropathy-related symptoms dur-
54 MALAGELADA ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 1, No. 1

Table 1. Comparison of Sucralfate-Treated and Placebo- which ulcer prophylaxis should be especially considered
Treated Patients at the End of the Study are subjects with chronic rheumatic disorders, advanced
Sucralfate Placebo age, taking NSAIDs continuously for more than 2
Parameter no. (%) no. (%) P value months, or giving high doses of the anti-inflammatory
Gastric mucosa, Lanza grade 0.0225 agent6,29 –31 with positive FOB test and gastric erosions,
0 17 (68) 8 (35)
preferably with history of peptic ulcer30,32,33 as well as
1 1 (4) 3 (13)
2 7 (28) 9 (39) with concomitant treatment with corticosteroids, anti-
3 0 2 (9) coagulants, or methotrexate.33,34
4 0 1 (4) Prior approaches to preventing the side effects of
5 0 0
Sum 25 (100) 23 (100) NSAIDs have included treatment with H2-receptor an-
Duodenal mucosa, Lanza grade 0.34 tagonist to inhibit acid secretion and the use of prosta-
0 24 (96) 20 (87) glandin analogues to replace the depleted endogenous
1 1 (4) 3 (13)
2 0 0
prostaglandins. However, H2-receptor antagonists are
3 0 0 not very effective for healing gastric erosions or prevent-
4 0 0 ing NSAID-associated gastric ulcers in patients who
5 0 0
Sum 25 23 (100) 0.188
continue ingestion of NSAIDs,35 and prostaglandin an-
Fecal occult blood test alogues, such as misoprostol, are limited by annoying
Positive 1 (4) 5 (21) gastrointestinal side effects such as diarrhea and abdom-
Negative 22 (96) 19 (79) inal cramps. Although proton-pump inhibitors have
Sum 23 (100) 24 (100)
been advocated to overcome poor tolerability of prosta-
glandin analogues, a recent study has shown that gastric
erosions healed significantly better in patients given
ing the study period were not significantly different misoprostol than in those treated with either 20 or 40
between both treatment groups both in regard to each mg of omeprazole.36
individual symptom and in regard to the total symptom Rationale for the use of sucralfate in NSAID compli-
score (sum of the scores of the 5 symptoms) (Table 2). cations are based on putative mechanisms of increase in
The lesions in the duodenal mucosa at the end of the cytoprotection and mucosal surface restitution and re-
study as measured with the Lanza scale was also similar duction of peptic and fibrinolytic activity. In this con-
among sucralfate- and placebo-treated patients (P ⫽ text, it is important to differentiate prophylaxis of either
0.34). acute or chronic NSAID lesions. In a double blind con-
Adverse Events trolled study on the effect of 1 g sucralfate 4 times daily
on gastric prostaglandin formation and microbleeding in
No instance of serious gastroduodenal complica- the intact and aspirin (2.5 g) treated stomach in healthy
tion (ulcer, hemorrhage, perforation) occurred after 42 male volunteers, both gastric microbleeding and DNA
days of treatment. Adverse events recorded in 5 patients loss were significantly lower in subjects given sucralfate
included diarrhea in 3 patients given sucralfate, rectal than in those treated with placebo.37 The protective
bleeding in 1 patient given placebo, and increased serum
effects of sucralfate on spontaneous gastric microbleeding
iron concentration in another patient treated with pla-
were accompanied by increased mucosal biosynthesis and
cebo.
luminal release of PGE2 6-keto-PGF1 alpha with a re-
duction in release of thromboxane B2.37 Other studies
Discussion have confirmed the favorable effect of sucralfate on aspi-
Research on prevention of gastrointestinal bleed- rin-induced gastric mucosal damage and acute NSAID-
ing associated with chronic NSAID therapies is highly
relevant because long-term NSAID administration fre-
Table 2. Total Symptom Scores During the Study
quently used for the treatment of arthritis and musculo-
skeletal disorders is well recognized to produce such Total symptom score (⫾SD)

complications.27 In fact, mucosal damage has been dem- Sucralfate Placebo P

onstrated in 50%–75% of arthritic patients under long-
Run-in 7.2 (3.93) 5.5 (3.50) 0.20
term NSAID therapy.28 Although a number of factors Baseline 7.9 (3.65) 5.5 (3.79) 0.07
can modify the risk of major gastrointestinal complica- 2 weeks 3.8 (2.13) 3.7 (3.22) 0.53
tions with NSAIDs and need to be considered when 4 weeks 3.2 (3.83) 3.2 (3.20) 0.89
6 weeks 2.1 (3.82) 2.5 (2.31) 0.23
individual prescribing decisions are made, candidates in
January 2003 CYTOPROTECTION OF SUCRALFATE 55

associated mucosal injury.26,33,38 – 40 Nevertheless, actual After a 6-week treatment period with 2 g (10 mL)
experience with the prophylactic effect of sucralfate on sucralfate twice daily, the percentage of patients with no
gastric mucosal lesions in chronic NSAID treatment is endoscopic gastric lesions was 68% compared with 35%
surprisingly limited.24 –26 It is well known that gastric in those treated with placebo. In this target population
erosions can progress to ulcers and can themselves lead to selected according to a marker of NSAID-related damage
gastrointestinal bleeding. Approximately, between 25% (positive FOB test) and using the most reliable diagnos-
and 50% of chronic NSAID users develop gastric ero- tic means of injury (endoscopy), chronic administration
sions, and their presence seems to be associated with a of sucralfate significantly improved NSAID-induced gas-
high risk of subsequently developing a gastric ulcer.41 tric erosions. Although a direct connection between
Several significant and distinctive features of our study treatment of erosions in NSAID-related gastropathy and
design deserve to be emphasized: Firstly, the use of the bleeding complications has not been established, we
believe that further studies on clinical use of sucralfate in
FOB test to screen for bleeding, which was always con-
patients receiving chronic NSAID therapy are warranted.
firmed by endoscopy. In a prospective study, Wroblewski
and Ostberg42 assessed the presence of upper gastroin- References
testinal disorders in a series of 170 geriatric inpatients 1. Garcı́a Rodrı́guez LA, Murray FE, Hawkey CJ, Yeomans ND. NSAID-
with positive FOB test. The prevalence of gastroduode- associated ulcers: new directions in research and clinical man-
nal ulcer confirmed by endoscopy or at autopsy was 22% agement. Oxford: Oxford Clinical Communications, 1996.
2. Committee of Safety of Medicines. Non-steroidal anti-inflamma-
and the sensitivity of the FOB test was 85%. Signifi- tory drugs and serious gastrointestinal adverse reactions. BMJ
cantly higher proportions of previous ulcer, iron defi- (Clin Res Ed) 1986;292:1190 –1191.
3. Fries JR, Miller SR, Spitz PW, Williams CA, Hubert HB, Bloch DA.
ciency anemia, and NSAID and steroid users were found
Towards an epidemiology of gastropathy associated with NSAIDs
among ulcer patients than among patients without ul- drug use. Gastroenterology 1989;96(suppl):649 – 655.
cerative upper gastrointestinal lesions. We are aware that 4. Hazleman BL. Incidence of gastropathy in destructive arthropa-
thies. Scand J Rheumatol 1989;78(suppl):1– 4.
many patients with gastric mucosal lesions may have 5. Castleden CM, Pickles H. Suspected adverse drug reactions in
been refused entry into the study because of a negative elderly patients reported to the Committee of Safety of Medi-
FOB pre-entry screening. But, conversely, by our ap- cines. Br J Clin Pharmacol 1988;26:347–353.
6. Griffin MR, Piper JM, Daugherty JR, Snowden M, Ray WA. Nonste-
proach we ensured that only patients with definite bleed- roidal anti-inflammatory drug use and increased risk for peptic
ing lesions would be randomized to the sucralfate or ulcer disease in elderly persons. Ann Intern Med 1991;114:257–
placebo arms of the study. A number of previous studies 263.
7. Larkai EN, Smith JL, Lisdky MD, Graham DY. Gastroduodenal
on NSAID-adjuvant preventive trials may have been mucosa and dyspeptic symptoms in arthritic patients during
inconclusive because entry was based on symptoms. chronic non-steroidal anti-inflammatory drug use. Am J Gastroen-
Symptoms are unreliable indicators in the context of terol 1987;82:1153–1158.
8. Mellem H, Stave R, Myren J, Osnes M, Hanssen LE, Mosvold J,
NSAID treatment for the identification of gastric lesions. Hebnes KL. Symptoms in patients with peptic ulcer and haemate-
Symptoms do not predict the presence of mucosal dam- mesis and/or melaena related to the use of NSAIDs. Scand J
age in NSAID-treated patients in whom the stomach and Gastroenterol 1985;82:1153–1158.
9. Abramson SB, Weissmann G. The mechanism of action of non-
duodenum are completely normal at endoscopy have steroidal anti-inflammatory drugs. Arthritis Rheum 1989;32:1–9.
been reported in 50%– 62% of the cases.7,24 –26,43 Thus, 10. Armstrong CP, Blower AL. NSAIDs and life-threatening complica-
in contrast to the design of previous studies,24 –26 in our tions of peptic ulceration. Gut 1987;28:527–532.
11. Kurata JH. An assessment of NSAIDs as a risk factor in ulcer
study the presence of gastrointestinal complaints was not diseases. Ann Intern Med 1991;114:311–317.
an inclusion criterion for the recruitment of patients. 12. Lanza FL, Graham DY, Davies RE, Rack MF. Endoscopic compar-
ison of cimetidine and sucralfate for prevention of naproxen-
Our entry criteria requiring patients treated with
induced acute gastroduodenal injury. Effect of scoring method.
NSAIDs for more than 2 months also prevented the Dig Dis Sci 1990;35:1494 –1499.
phenomenon of “gastric adaptation” in which visible 13. Butt JH, Barthel JS, Moore RA. Clinical spectrum of upper gas-
trointestinal effects of NSAIDs. Am J Med 1988;84(suppl 2A):5–
early gastric mucosal injury lessens or resolves com- 14.
pletely within 12 hours to 24 days despite continued 14. Bjilsma JWJ. Multicenter Group. Treatment of endoscopy nega-
administration of the offending agent.44,45 Finally, the tive NSAID-induced upper gastrointestinal symptoms with cime-
tidine: an international multicenter collaborative study. Aliment
profile of patients included in the study took into ac- Pharmacol Ther 1988;25:75– 83.
count different risk factors30,33 (e.g., alcohol ingestion, 15. Hollander D. Gastrointestinal complications of NSAIDs: prophy-
potential eradication of H. pylori in the course of the trial, lactic and therapeutic strategies. Am J Med 1994;96:274 –281.
16. Ligumsky M, Karmeli F, Rachmilewitz D. Sucralfate protection
use of non-allowed concurrent medication, etc.) that may against gastrointestinal damage: possible role of prostanoids.
constitute a source of bias and should be avoided. Isr J Med Sci 1986;22:801– 806.
56 MALAGELADA ET AL. CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 1, No. 1

17. Hollander D, Tarnawski A, Gergely H, Zipser RD. Sucralfate pro- 33. Fries JF, Williams CA, Bloch DA, Michel BA. Nonsteroidal anti-
tection of the gastric mucosa against ethanol-induced injury: a inflammatory drug-associated gastropathy: incidence and risk
prostaglandin mediated process. Scand J Gastroenterol 1984; factor models. Am J Med 1991;91:213–222.
19(suppl 101):97–102. 34. Piper JM, Ray WA, Daugherty JR, Griffin MR. Costicosteroid use
18. Nexo E, Poulsen SS. Does epidermal growth factor play a role in and peptic ulcer disease: role of nonsteroidal anti-inflammatory
the action of sucralfate? Scand J Gastroenterol 1987;22(suppl drugs. Ann Intern Med 1991;114:735–740.
127):45– 49. 35. Yeomans ND, Tulassay Z, Juhász L, Rácz I, Howard JM, Van
19. Tarnawski A, Tanoue K, Santos AM, Sarfeh IJ. Cellular and mo- Rensburg CJ, Swannell AJ, Hawkey CJ. Comparison of omepra-
lecular mechanisms of gastric ulcer healing. Is the quality of zole with ranitidine for ulcers associated with nonsteroidal anti-
mucosal scar affected by treatment? Scand J Gastroenterol inflammatory drugs. N Engl J Med 1998;338:719 –726.
1995;210(suppl):9 –14. 36. Hawkey CJ, Karrasch JA, Szczepanski L, Walker DG, Barkun A,
20. Cohen MM, Bowdler R, Gervais P, Morris GP, Wang HR. Sucral- Swannell AJ, Yeomans ND. Omeprazole compared with misopros-
fate protection of human gastric mucosa against acute ethanol tol for ulcers associated with nonsteroidal antiinflammatory
injury. Gastroenterology 1989;92:292–298. drugs. N Engl J Med 1998;338:727–734.
21. Kaya N, Boyunapa H, Kahraman H, Altintop L. Protective effects 37. Konturek SJ, Kwiecien N, Obtulowicz W, Kopp B, Olesky J. Double
of sucralfate and omeprazole on gastric mucosal damage in- blind controlled study on the effect of sucralfate on gastric pros-
duced by ethanol in rats. Wien Klin Wochenschr 1998;110:96 – taglandin formation and microbleeding in normal and aspirin
100. treated man. Gut 1986;27:1450 –1456.
22. Tarnawski A, Hollander D, Krause WJ, Zipser RD, Stachura J, 38. Tesler MA, Lim ES. Protection of gastric mucosa by sucralfate
Gergely H. Does sucralfate affect the normal gastric mucosa? from aspirin-induced erosions. J Clin Gastroenterol 1981;3(suppl
Histological, ultrastructural, and functional assessment in the 2):175–179.
rat. Gastroenterology 1986;90:893–905. 39. Stern AI, Ward F, Hartley G. Protective effect of sucralfate against
23. Wallence JL, Morris GP, Beck PL, Williamson TE, Gingras GR. aspirin-induced damage to the human gastric mucosa. Am J Med
Effect of sucralfate on gastric prostaglandin and leukotriene 1987;83(suppl 3B):83– 85.
synthesis: relationship to protective actions. Can J Physiol Phar- 40. Shepherd HA, Fine D, Hillier K, Jewell R, Cox N. Effect of sucral-
macol 1988;66:666 – 670. fate and cimetidine on rheumatoid patients with active gastrodu-
24. Wu WC, Semble EL, Castell DO, Caldwell JR, Roth SH, Heller MD, odenal lesions who are taking NSAID. A pilot study. Am J Med
Marsh WH. Sucralfate therapy of nonsteroidal anti-inflammatory 1989;86:49 –54.
drug-induced gastritis (abstr). Gastroenterology 1985;88:1636. 41. Agrawal NM, Roth S, Graham DY, White RH, Germain B, Brown JA,
25. Wu WC, Semble EL, Castell DO, Caldwell JR, Roth SH, Heller MD, Stromatt SC. Misoprostol compared with sucralfate in the pre-
Marsh WH. Chronic sucralfate therapy of nonsteroidal anti-inflam- vention of nonsteroidal anti-inflammatory drug-induced gastric
matory drug-induced gastritis: a multi-center trial (abstr). Gastro- ulcer. A randomized, controlled trial. Ann Intern Med 1991;115:
enterology 1986;90:1696. 195–200.
26. Caldwell JR, Roth SH, Wu WC, Semble EL, Castell DO, Heller MD, 42. Wroblewski M, Ostberg H. Ulcer disease among geriatric inpa-
Marsh WH. Sucralfate treatment of NSAID-induced gastrointesti- tients with positive faecal occult blood test and/or iron deficiency
nal symptoms and mucosal damage. Am J Med 1987;83(Suppl anemia. A prospective study. Scand J Gastroenterol 1990;25:
3B):74 – 82. 489 – 495.
27. Agrawal NM, Aziz K. Prevention of gastrointestinal complications 43. Caruso I, Bianchi Porro G. Gastroscopic evaluation of anti-inflam-
associated with nonsteroidal antiinflammatory drugs. J Rheuma- matory agents. BMJ 1980;280:75–78.
tol 1998;51(suppl):17–20. 44. Olivero JJ, Graham DY. Gastric adaptation to NSAID drugs in man.
28. Lichtenstein DR, Syngal S, Wolfe MM. Nonsteroidal antiinflam- Scand J Gastroentrol 1992;27(suppl 193):53–58.
matory drugs and the gastrointestinal tract. The double-edged 45. Konturek JW, Dembinski A, Stoll R, Domschke W, Konturek SJ.
sword. Arthritis Rheum 1995;38:5–18. Mucosal adaptation to aspirin-induced gastric damage in hu-
29. Henry D, Dobson A, Turner C. Variability in the risk of major mans. Studies on blood flow, gastric mucosal flow, gastric mu-
gastrointestinal complications from nonaspirin nonsteroidal anti- cosal growth, and neutrophil activation. Gut 1994;35:
inflammatory drugs. Gastroenterology 1993;105:1078 –1088. 1197–1204.
30. Garcia Rodriguez LA, Jick H. Risk of upper gastrointestinal bleed-
ing and perforation associated with individual non-steroidal anti-
inflammatory drugs. Lancet 1994;343:769 –772. Address requests for reprints to: Juan R. Malagelada, M.D., Servicio
31. Laporte JR, Carne X, Vidal X, Moreno V, Juan J. Upper gastroin- de Aparato Digestivo, Hospital Universitari Vall d’Hebron, Passeig Vall
testinal bleeding in relation to previous use of analgesics and d’Hebron 119-129, 08035 Barcelona, Spain. e-mail: malagela@hg.
non-steroidal anti-inflammatory drugs. Catalan Countries Study vhebron.es.
on Upper Gastrointestinal Bleeding. Lancet 1991;337:85– 89. The authors thank Drs. H. Brinkhoff, P. Krupp, S. Kaspari, and A.
32. McCarthy DM. Nonsteroidal antiinflammatory drugs – the clinical Ansari for their contribution to the study and Dr. Marta Pulido for
dilemmas. Scand J Gastroenterol 1992; 27(suppl 192):9 –16. editing the manuscript and editorial assistance.