Blackwell Science, LtdOxford, UKJGHJournal of Gastroenterology and Hepatology0815-93192005 Blackwell Publishing Asia Pty LtdS1 200521S198102Original Article Endoscopy

and H. pylori in NSAID-GUT Kamada

et al.

Journal of Gastroenterology and Hepatology (2006) 21, 98102

DOI: 10.1111/j.1440-1746.2005.04219.x

GASTROENTEROLOGY

Endoscopic characteristics and Helicobacter pylori infection in NSAID-associated gastric ulcer

TOMOARI KAMADA,* JIRO HATA,* HIROAKI KUSUNOKI,* KUNIAKI SUGIU,* TATSURO TANIMOTO, MITSUHIRO MIHARA, HIROSHIGE HAMADA, SOICHIRO KIDO, QIAN DONGMEI AND KEN HARUMA*

*Division of Gastroenterology, Department of Internal Medicine, Kawasaki Medical School, Kurashiki, Department of Internal Medicine, Saiseikai Hiroshima Hospital, Department of Internal Medicine, Hiroshima Railway Hospital, Hiroshima, Department of Internal Medicine, Mitsubishi Mihara Hospital, Mihara, Japan and Digestive Department of Internal Medicine, Beijing Tongren Hospital, Capital University of Medical Sciences, Beijing, China

Abstract Background and Aim: Helicobacter pylori infection and non-steroidal anti-inammatory drugs (NSAIDs) are deeply involved in the etiology of gastric ulcers. The aim of our study was to clarify the endoscopic characteristics and H. pylori infection status of NSAID-associated gastric ulcers. Methods: The study group comprised 50 patients (23 men, 27 women; mean age 66.5 years) with NSAID-associated gastric ulcers and 100 sex- and age-matched patients with gastric ulcer associated with other factors (control group). Ulcer morphology, size and number of lesions, onset site and incidence of hemorrhagic ulcers were investigated endoscopically in both groups. H. pylori infection was diagnosed by serology, histology and 13C-urea breath test. Results: Multiple lesions (68% vs 20%, P < 0.001), occurrence in the antrum (56% vs 6%, P < 0.001), and hemorrhagic ulcer (34% vs 4%, P < 0.001) were signicantly more prevalent in patients with NSAID-associated gastric ulcers than in patients with non-NSAID-associated gastric ulcer. The H. pylori infection rate was signicantly lower in NSAID-associated gastric ulcer patients than in nonNSAID-associated gastric ulcer patients (48% vs 96%, P < 0.001). In the NSAID-associated gastric ulcer group, the prevalence of H. pylori infection was signicantly lower in patients with ulcers in the antrum than in those with ulcers in the angulus or corpus (25% vs 77.3%, P < 0.001). Conclusions: In contrast to non-NSAID-associated gastric ulcers, NSAID-associated gastric ulcers frequently occur in the antrum with bleeding. The rate of H. pylori infection in NSAID-associated gastric ulcers is signicantly lower than that in non-NSAID-associated gastric ulcers. 2005 Blackwell Publishing Asia Pty Ltd Key words: endoscopy, gastric ulcer, Helicobacter pylori, hemorrhagic ulcer, non-steroidal antiinammatory drugs.

INTRODUCTION
There is now strong evidence that Helicobacter pylori infection plays an important role in the pathogenesis of chronic gastritis, peptic ulcers and gastric carcinoma.13 Additionally, H. pylori infection and nonsteroidal anti-inammatory drugs (NSAIDs) are considered the two major causes of peptic ulcer diseases,

with NSAIDs being the major cause of H. pylorinegative peptic ulcer disease.46 Chronic administration of NSAIDs is associated with an increased incidence of signicant adverse events such as upper gastrointestinal hemorrhage or perforation.79 According to epidemiological studies, the risks of peptic ulcer and death are three to six times higher among people who take these drugs than among those who do not.10

Correspondence: Tomoari Kamada, Division of Gastroenterology, Department of Internal Medicine, Kawasaki Medical School, Matsushima 577, Kurashiki, 701-0192, Japan. Email: tkamada@med.kawasaki-m.ac.jp Accepted for publication 12 July 2005.

Endoscopy and H. pylori in NSAID-GU Non-steroidal anti-inammatory drugs comprise one of the most commonly used classes of medication worldwide. Singh et al.11 reported that NSAIDs are regularly used by at least 13m people: 2m with rheumatoid arthritis (RA), 3m with probable RA, and 8m with osteoarthritis. It is conservatively estimated that 16 500 NSAID-related deaths occur among these patients every year in the United States.11 In Japan, the incidence of gastric ulcer (GU) attributable to NSAIDs is expected to increase as the population ages. The aim of our study was to clarify the endoscopic characteristics and H. pylori infection status of NSAIDassociated GU by comparing them to the endoscopic characteristics and H. pylori infection status of nonNSAID-associated GU in Japanese patients.

99

Endoscopy and biopsy

After providing informed consent, all patients underwent endoscopy and biopsy after an overnight fast without premedication. GU was diagnosed when an open crater in the gastric mucosa was veried endoscopically. We ruled out malignant ulcer by means of gastric biopsy at the ulcers edge. At endoscopy, ulcer morphology (round or irregular), size (less than 10 mm, 1020 mm, more than 20 mm), number of lesions (multiple or single), onset site (antrum, angulus, corpus) and incidence of hemorrhagic ulcer were investigated in both groups. We considered the ulcer morphology to be irregular unless the ulcer appeared round or oval. We estimated the size of the ulcer by placing biopsy forceps alongside the ulcer; the fully opened cup of the forceps was 5 mm in diameter. In patients with multiple ulcers, the location of the largest ulcer was recorded. Hemorrhagic ulcer was diagnosed in patients who had upper gastrointestinal bleeding of Forrest class12 Ia, Ib or IIa. Three biopsy specimens were obtained from each patient, one from the greater curvature of the midantrum, one from the lesser curvature of the angulus and one from the greater curvature of the mid-corpus. Specimens were xed in 10% buffered formalin, embedded in parafn, cut into 4-m sections, and stained with Giemsa stain for H. pylori identication. All Giemsa-stained biopsy specimens were reviewed by two pathologists who were blinded to the subjects symptoms and laboratory data. The prevalence of H. pylori infection was investigated in both groups.

PATIENTS AND METHODS

Patients
The study was conducted at the Division of Gastroenterology, Department of Internal Medicine, Kawasaki Medical School from October 2000 through December 2002. The study subjects comprised 50 patients with NSAID-associated GU (23 men, 27 women; mean age 66.5 years, range 2581 years) and 100 sex- and agematched patients with non-NSAID-associated GU (control group; 46 men, 54 women; mean age 66.5 years, range 2578 years). NSAID-associated GU was dened as GU diagnosed on the basis of endoscopic examination in patients who had received treatment with NSAIDs, regardless of the presence of H. pylori infection. NSAIDs were used to treat lumbago (26 patients), old cerebral infarction (six patients), ischemic heart disease (ve patients), RA (four patients) and other conditions (nine patients). Chief complaints in the NSAID-associated GU group were epigastric pain or discomfort (23 patients), upper gastrointestinal hemorrhage (15 patients) and other symptoms (two patients). Ten patients were asymptomatic. The NSAIDs prescribed were low-dose aspirin (12 patients), loxoprofen (10 patients), diclofenac (nine patients), lornoxicam (six patients), indomethacin (ve patients) and others (eight patients). Thirty-six patients (72%) were chronic NSAID users, and 14 (28%) were on-demand NSAID users. We dened chronic NSAID users as patients who had received treatment with NSAIDs for more than 1 month. Patients in the control group had no history of NSAID treatment, and GU in these patients was diagnosed on the basis of endoscopic examination mainly for abdominal symptoms. Patients with malignancy, previous gastric surgery or a history of H. pylori eradication therapy and patients currently taking antibiotics, proton-pump inhibitors, H2-receptor antagonists or misoprostol were excluded from the study. All patients provided informed consent before undergoing endoscopic examination. The study protocol was approved by the Ethics Committee of the gastrointestinal unit at Kawasaki Medical School.

Assessment of H. pylori infection

Immunoglobulin G (IgG) antibodies to H. pylori were measured by enzyme-linked immunosorbent assay (HEL-p TEST; AMRAD, Kew, Victoria, Australia). We categorized a subject as serologically negative if the titer was below 30 units/mL (sensitivity, 100%; specicity, 88.9%), as previously described.13 The 13C-urea breath test (UBT) was performed after an overnight fast. The result was considered positive if 13CO2 excretion exceeded 2.5 per thousand, as we previously described (sensitivity, 100%; specicity, 96%).14 H. pylori infection status was determined by histology, serology and UBT. Subjects were considered to be infected with H. pylori if at least one of these three methods yielded a positive result. Subjects were considered to be uninfected if all three methods yielded a negative result.

Statistical analysis
Results are expressed as percentages. Between-group differences were evaluated by the MannWhitney Utest, and P < 0.05 was considered signicant.

RESULTS
Baseline patient characteristics in the NSAIDassociated GU group and non-NSAID-associated GU

100
Table 1 Baseline patient characteristics in the two study groups NSAID-associated GU (n = 50) Mean age (range, year) Gender (M/F) Smoking history (+) Ulcer history (+) 66.5 (2581) 23/27 12 (24%) 9 (18%)

T Kamada et al.

non-NSAID-associated GU (n = 100) 66.5 (2578) 46/54 21 (21%) 23 (23%)

P-value

0.67 0.48

GU, gastric ulcer; NSAID, non-steroidal anti-inammatory drug. Table 2 Endoscopic characteristics in the two study groups NSAID-associated GU (n = 50) Morphology Round Irregular Size <10 mm 1020 mm >20 mm Number of lesions Multiple Single Ulcer site Antrum Angulus Corpus Hemorrhagic ulcer H. pylori infection non-NSAID-associated GU (n = 100) P-value

39 (78%) 11 (22%) 20 (40%) 22 (44%) 8 (16%) 34 (68%) 16 (32%) 28 5 17 17 24 (56%) (10%) (34%) (34%) (48%)

82 (82%) 18 (18%) 42 (42%) 48 (48%) 10 (10%) 20 (20%) 80 (80%) 6 32 62 4 96 (6%) (32%) (62%) (4%) (96%)

0.55

0.56

<0.001

<0.001

<0.001 <0.001

GU, gastric ulcer; NSAID, non-steroidal anti-inammatory drug.

group appear in Table 1. There were no signicant differences in age, gender, smoking or ulcer recurrence between the two groups. The NSAID-associated GU group had a signicantly higher prevalence of multiple lesions in comparison to that of the non-NSAID-associated GU group (34/50, 68% vs 20/100, 20%; P < 0.001). The ulcerating lesions were located in the antrum, angulus and corpus of the stomach in 56%, 10% and 34%, respectively, of the NSAID-associated GU patients and in 6%, 32% and 62%, respectively, of the non-NSAID-associated GU patients. Antral onset (28/50, 56% vs 6/100, 6%; P < 0.001) and hemorrhagic ulcer (17/50, 34% vs 4/ 100, 4%; P < 0.001) were signicantly more prevalent in the NSAID-associated GU group than in the nonNSAID-associated GU group. There was no difference in ulcer morphology or size between the two groups (Table 2). Patients in the NSAID-associated group were further classied, according to the type of NSAID used, into a low-dose aspirin-associated GU group and a non-low-dose aspirin-associated GU group. Round ulcers were more prevalent in the non-low-dose aspirinassociated GU group than in the low-dose aspirinassociated GU group (30/38, 78.9% vs 9/12, 75%; P = 0.78), but the difference was not statistically signicant.

Table 3 Prevalence of H. pylori infection according to ulcer site in patients with NSAID-associated GU No. of patients 28 22 No. of H. pylori-positive patients (%) 7 (25)* 17 (77.3)

Ulcer site Antrum Angulus or corpus

*P < 0.001. GU, gastric ulcer; NSAID, non-steroidal anti-inammatory drug.

The overall rate of H. pylori infection was signicantly lower in the NSAID-associated GU group than in the non-NSAID-associated GU group (24/50, 48% vs 96/ 100, 96%; P < 0.001) (Table 2). In the NSAIDassociated GU group, the prevalence of H. pylori infection was signicantly lower in patients with ulcers in the antrum than in those with ulcers in the corpus or angulus (7/28, 25% vs 17/22, 77.3%; P < 0.001) (Table 3). Furthermore, antral onset was signicantly more common in H. pylori-negative cases than in H. pyloripositive cases (H. pylori-negative: 21/26, 80.7% vs H. pylori-positive: 7/24, 29.2%; P < 0.001), whereas

Endoscopy and H. pylori in NSAID-GU

101

Figure 1 Endoscopic examination shows non-steroidal antiinammatory drugs (NSAID)-associated multiple hemorrhagic ulcers with exposed blood vessels in the antrum of the stomach, which was negative for Helicobacter pylori infection and without atrophic gastritis.

Figure 2 Endoscopic examination shows NSAIDassociated multiple ulcers in the corpus of the stomach, which was positive for H. pylori infection and showed atrophic gastritis.

corporeal onset was signicantly more common in H. pylori-positive cases than in H. pylori-negative cases (H. pylori-positive: 14/24, 58.3% vs H. pylori-negative: 3/26, 11.5%; P < 0.001). In addition, the rate of H. pylori infection was lower in the non-low-dose aspirin-associated GU group than in the low-dose aspirin-associated GU group (17/38, 44.7% vs 7/12, 58.3%; P = 0.41), but the difference was not statistically signicant. Endoscopic ndings are presented in Figures 1 and 2.

DISCUSSION
Our ndings indicate that the majority of NSAIDassociated GU develop in the antrum with hemorrhage, in comparison to non-NSAID-associated GU, and that the rate of H. pylori infection in patients with NSAIDassociated GU, especially antral ulcer, is signicantly lower than that in patients with non-NSAID-associated GU. We found a signicantly higher prevalence of multiple lesions in the antrum with upper or lower gastrointestinal tract bleeding in patients with NSAID-associated GU than in those with non-NSAID-associated GU. However, there was no difference in ulcer morphology or size between the two groups. Mizokami et al.15 reported that most NSAID-associated GU develop in the antrum, unlike ordinary GU, which commonly occur in the corpus or angulus of the stomach. Other studies16,17 have also shown that NSAID-associated GU occur mainly in the antrum, whereas the majority of non-NSAID-associated GU occur in the corpus. al-Assi et al.9 reported a high frequency of NSAID use in patients presenting with upper gastrointestinal tract bleeding and noted that 74% of patients with upper gas-

trointestinal hemorrhage from peptic ulcer were taking NSAIDs. In our study, the overall rate of H. pylori infection was signicantly lower in the NSAID-associated GU group than in the non-NSAID-associated GU group. Other studies have also shown a lower prevalence of H. pylori infection in patients with NSAID-associated GU than in patients with non-NSAID-associated GU.14,1820 In a study of 217 patients, Ng et al.20 found a lower prevalence of H. pylori in GU patients taking NSAIDs than in those not taking them. In the present study, we investigated the relation between the ulcer site and H. pylori infection in patients with NSAID-associated GU. We found that the rate of H. pylori infection was signicantly lower in patients with onset in the antrum than in those with onset in the corpus or angulus in the NSAID-associated GU group. In other words, antral onset was signicantly more common in the H. pylorinegative patients than in the H. pylori-positive patients, whereas corporeal onset was signicantly more common in the H. pylori-positive patients than in the H. pylori-negative patients. Mizokami et al.15 also reported that the prevalence of H. pylori infection was signicantly lower in patients with an NSAIDassociated GU in the antrum than in those with an ulcer in the corpus or angulus of the stomach. Thus, GU in the antrum are associated more strongly with NSAID use than with H. pylori infection, whereas GU in the corpus and angulus are associated more strongly with H. pylori infection than with NSAID use. In Japan, H. pylori infection commonly induces atrophic gastritis;1 it also decreases gastric acid secretion with age.21 Thus, H. pylori infection may play a partially protective role against the development of NSAID-associated GU by decreasing gastric acid secretion. This could explain why the rate of H. pylori infection was lower with

102 NSAID-associated GU than with non-NSAIDassociated GU in our study. In addition, our results suggest that GU develop in the corpus according to the progress of atrophic gastritis in H. pylori-positive patients, whereas GU develop in the antrum without atrophic gastritis in H. pylori-negative patients. In conclusion, our ndings indicate that NSAIDassociated GU frequently occurs in the antrum with bleeding in comparison to non-NSAID-associated GU and that the rate of H. pylori infection in patients with NSAID-associated GU is signicantly lower than that in patients with non-NSAID-associated GU.

T Kamada et al.
and nonusers of nonsteroidal anti-inammatory drugs. Endoscopy 2003; 35: 32732. al-Assi MT, Genta RM, Karttunen TJ, Graham DY. Ulcer site and complications: relation to Helicobacter pylori infection and NSAID use. Endoscopy 1996; 28: 22933. Barrier CH, Hirschowitz BI. Controversies in the detection and management of nonsteroidal antiinammatory drug-induced side effects of the upper gastrointestinal tract. Arthritis Rheum. 1989; 32: 92632. Singh G, Triadalopoulos G. Epidemiology of NSAID induced gastrointestinal complications. J. Rheumatol. 1999; 26 (Suppl. 56): S1824. Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in gastrointestinal bleeding. Lancet 1974; 2: 3947. Midolo PD, Lambert JR, Russell EG, Lin SK. A practical single sample dry latex agglutination test for Helicobacter pylori antibody detection. J. Clin. Pathol. 1995; 48: 96971. Chen X, Haruma K, Kamada T et al. Factors that affect results of the 13C urea breath test in Japanese patients. Helicobacter 2000; 5: 98103. Mizokami Y, Narushima K, Shiraishi T, Otsubo T, Narasaka T, Matsuoka T. Non-Helicobacter pylori ulcer disease in rheumatoid arthritis patients receiving longterm NSAID therapy. J. Gastroenterol. 2000; 35 (Suppl. 12): S3841. Waki S, Kinoshita Y, Fukui H et al. Intragastric distribution of nonsteroidal anti-inammatory drug-related ulcers in patients without collagen diseases. J. Clin. Gastroenterol. 1997; 25: 5924. Cheatum DE, Arvanitakis C, Gumpel M et al. An endoscopic study of gastroduodenal lesions induced by nonsteroidal anti-inammatory drugs. Clin. Ther. 1999; 21: 9921003. Mizokami Y, Tamura K, Fukuda Y, Yamamoto I, Shimoyama T. Non-steroidal anti-inammatory drugs associated with gastroduodenal injury and Helicobacter pylori. Eur. J. Gastroenterol. Hepatol. 1994; 6 (Suppl. 1): S10912. Matsukawa Y, Aoki M, Nishinarita S et al. Prevalence of Helicobacter pylori in NSAID users with gastric ulcer. Rheumatology (Oxford) 2003; 42: 94750. Ng TM, Fock KM, Khor JL et al. Non-steroidal antiinammatory drugs, Helicobacter pylori and bleeding gastric ulcer. Aliment. Pharmacol. Ther. 2000; 14: 2039. Haruma K, Kamada T, Kawaguchi H et al. Effect of age and Helicobacter pylori infection on gastric acid secretion. J. Gastroenterol. Hepatol. 2000; 15: 27783.

10

11

12 13

REFERENCES
1 Kawaguchi H, Haruma K, Komoto K, Yoshihara M, Sumii K, Kajiyama G. Helicobacter pylori infection is the major risk factor for atrophic gastritis. Am. J. Gastroenterol. 1996; 91: 95962. 2 Mihara M, Haruma K, Kamada T et al. The role of endoscopic ndings for the diagnosis of Helicobacter pylori infection: evaluation in a country with high prevalence of atrophic gastritis. Helicobacter 1999; 4: 408. 3 Komoto K, Haruma K, Kamada T et al. Helicobacter pylori infection and gastric neoplasia: correlations with histological gastritis and tumor histology. Am. J. Gastroenterol. 1998; 93: 12716. 4 Borody TJ, Brandl S, Andrews P, Jankiewicz E, Ostapowicz N. Helicobacter pylori-negative gastric ulcer. Am. J. Gastroenterol. 1992; 87: 14036. 5 Nishikawa K, Sugiyama T, Kato M et al. Non-Helicobacter pylori and non-NSAID peptic ulcer disease in the Japanese population. Eur. J. Gastroenterol. Hepatol. 2000; 12: 63540. 6 Kamada T, Haruma K, Kusunoki H et al. Signicance of an exaggerated meal-stimulated gastrin response in pathogenesis of Helicobacter pylori-negative duodenal ulcer. Dig. Dis. Sci. 2003; 48: 64451. 7 Hernandez-Diaz S, Rodriguez LA. Association between nonsteroidal anti-inammatory drugs and upper gastrointestinal tract bleeding/perforation: an overview of epidemiologic studies published in the 1990s. Arch. Intern. Med. 2000; 160: 20939. 8 Adamopoulos A, Efstathiou S, Tsioulos D, Tsami A, Mitromaras A, Mountokalakis T. Acute upper gastrointestinal bleeding: comparison between recent users 14

15

16

17

18

19

20

21