Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Helicobacter pylori is associated with gastritis, duodenal ulcer, and gastric cancer

ACP J Club. 1994 May-June;120:62. doi:10.7326/ACPJC-1994-120-3-062

Source Citations

Veldhuyzen van Zanten SJ, Sherman PM. Helicobacter pylori infection as a cause of gastritis, duodenal ulcer, gastric cancer and nonulcer dyspepsia: a systematic overview. Can Med Assoc J. 1994 Jan 15;150:177-85.

Veldhuyzen van Zanten SJ, Sherman PM. Indications for treatment of Helicobacter pylori infection: a systematic overview. Can Med Assoc J. 1994 Jan 15;150:189-98.



To evaluate the association between Helicobacter pylori infection and gastritis, duodenal ulcer, nonulcer dyspepsia, and gastric cancer; the treatment options for eradication of H. pylori; and whether eradication of H. pylori is indicated in patients with duodenal ulcer, nonulcer dyspepsia, and gastric cancer.

Data sources

English-language studies were identified in the MEDLINE databases (1983 to December 1992) using key words Helicobacter pylori, gastritis, duodenal ulcer, gastric cancer, dyspepsia, and clinical trial; a review of Current Contents; a hand search of 6 journals (New England Journal of Medicine, Lancet, Gastroenterology, Gut, Digestive Diseases and Sciences, and American Journal of Gastroenterology); and bibliographies of relevant studies and review articles.

Study selection

Studies were selected if they included at least 25 patients. Case reports and review articles were also evaluated. Papers investigating cause were further screened for quality and strength of evidence. Therapy papers were further screened to include randomized controlled trials that compared anti- H. pylori therapy in adults with conventional therapy (8 trials). For nonulcer dyspepsia, only placebo-controlled trials were included (6 trials).

Data extraction

All studies were scored for quality on a 4-point scale (high-quality methodology with no major weaknesses, reasonable-quality methodology with some weaknesses, weak study with definite shortcomings, and poor study).

For therapy articles, data were extracted on disorder (e.g., duodenal or nonulcer dyspepsia), treatments used, duration of treatment, numbers of treated patients, rate of ulcer healing, rate of H. pylori eradication, and rate of relapse for disease and symptoms.

Each article on causation was analyzed for disorder studied, study design (review, randomized controlled trial, cohort study, case series, case report), and strength of evidence for association on a 6-point scale (strong, moderate, or weak support, causal decision unaffected, and moderate or weak evidence against) for each of 8 established guidelines to determine causation (evidence from experiments in humans, strong association, consistent association across studies, temporal relation, dose-response gradient, epidemiologic sense, biologic sense, and specific association).

Main results

No data were pooled because of study heterogeneity. For the association between H. pylori and duodenal ulcer, nonsteroidal anti-inflammatory drug-induced duodenal ulcers were excluded. Strong evidence indicated that H. pylori was a causal factor in both gastritis (histologic evidence of mucosal inflammation in the stomach) and duodenal ulcer formation, although no direct evidence existed in the evaluated studies that H. pylori infection in humans preceded the development of duodenal ulcer. Moderate epidemiologic evidence supported a causal relation between H. pylori and gastric cancer, including both intestinal and diffuse-type gastric adenocarcinomas and gastric lymphoma. The studies included no evidence for an association between H. pylori and nonulcer dyspepsia.

Monotherapy eradication rates for H. pylori reached 20% with bismuth and 23% with amoxicillin. Dual-therapy eradication rates reached 55% using bismuth, amoxicillin, and metronidazole and 81% using omeprazole and amoxicillin. Triple therapies with bismuth, metronidazole and either amoxicillin or tetracycline were the most successful (73% to 94%). Adding anti- H. pylori treatment to conventional therapy healed ulcers more quickly. Duration of treatment was usually 2 weeks. Up to 45% of patients on dual or triple therapy had side effects, usually minor. The relapse rates for duodenal ulcer after triple therapy at 1 year were < 10% compared with 60% to 80% for conventional acid-suppressive therapy. For nonulcer dyspepsia, study quality was poor and no evidence in the studies supported anti- H. pylori treatment in patients with nonulcer dyspepsia who were H. pylori positive. In the studies evaluated, no data existed on the effect of eradication of H. pylori on subsequent risk for gastric cancer.


The evidence supports a strong causal relation between Helicobacter pylori infection and gastritis and duodenal ulcer and a moderate association with gastric cancer. No evidence is presented for a causal relation between Helicobacter pylori and nonulcer dyspepsia.

Treatment of adults who have duodenal ulcer disease and Helicobacter pylori infections with anti- Helicobacter pylori agents decreases treatment times and reduces ulcer relapse. Triple therapy is better than monotherapy. No evidence in the studies supports therapy for nonulcer dyspepsia.

Source of funding: No external funding.

For article reprint: Dr. S.J.O. Veldhuyzen van Zanten, QEII, VGH Site, Room 928, South Wing, Centennial Bldg., Halifax, Nova Scotia B3H 2Y9, Canada. FAX 902-473-4408.


The long-held hypothesis that duodenal ulcer disease is caused primarily by acid has, after a decade of siege by the H. pylori hypothesis, finally collapsed. That the acid hypothesis could even be challenged, much less toppled, appeared as unthinkable 10 years ago as the fall of Communism in the former USSR. Within the last few years, strong evidence has accumulated, however, about H. pylori's importance, persuading even this previously skeptical writer (1). Clinical practice will be dramatically changed by the evidence summarized by Veldhuyzen van Zanten and Sherman and by the recent recommendations of an NIH Consensus Conference (2).

First, H. pylori infection can now be considered the most common cause of duodenal ulcer, with nonsteroidal anti-inflammatory drugs and, rarely, the Zollinger-Ellison syndrome as other causes. The Consensus Conference reviewed the evidence, as did Veldhuyzen van Zanten, and made several important recommendations about therapy:

1. Treatment for 2 weeks with triple antibiotics (bismuth subsalicylate, tetracycline, and metronidazole) yields eradication rates of about 90%, whereas substitution of amoxicillin for tetracycline or metronidazole only slightly lowers efficacy.

2. "All patients with gastric or duodenal ulcers who are infected with H. pylori should be treated with antimicrobial agents regardless of whether they are suffering from the initial presentation of the disease or from a recurrence." (2). This recommendation—to treat all ulcers (i.e., not just failures of histamine-2 therapy, as some had earlier recommended)—signals a new aggressiveness toward the role of antimicrobial agents in the primary care of ulcers.

3. "Peptic ulcer patients who are infected with H. pylori and are undergoing maintenance treatment with antisecretory agents should also be treated for the infection." Thus, persons with previously diagnosed ulcers currently on maintenance should now receive antimicrobial therapy.

Veldhuyzen van Zanten and Sherman and the Consensus Conference are cautious about concluding a cause-and-effect role of H. pylori for nonulcer dyspepsia. It seems likely, however, that at least some persons with nonulcer dyspepsia may have symptoms caused by H. pylori, simply because extensive experience from clinical trials of histamine-2 blockers shows that persons who have had ulcers may have "ulcer symptoms" even after their ulcers have healed. Presumably, some of these persons have dyspepsia related to ulcers that would respond to therapy. The clinical challenge, of course, is to describe and classify persons with "dyspepsia" in a way that helps illuminate specific cause and the likelihood of response to therapy. Further randomized controlled trials comparing anti- H. pylori treatment with placebo for symptomatic relief of symptoms are needed.

Regarding gastric cancer, both Veldhuyzen van Zanten and Sherman and the Consensus Conference conclude that an association with H. pylori exists but with current understanding, the cancer risk is not high enough to warrant therapy.

The H. pylori story will continue to unfold. For now, we have major new recommendations, based on strong evidence, that H. pylori should be treated in persons with ulcers. We may speculate that interest in identifying and treating even asymptomatic infected persons will occur—if both screening tests and therapy can be made inexpensive and easy—to prevent ulcers and, perhaps, cancer.

In the meantime, the evidence is sufficient to dramatically change practice and our way of thinking about pathogenesis!

David F. Ransohoff, MD
University of North CarolinaChapel Hill, North Carolina, USA


1. Rabeneck L, Ransohoff DF. Is Helicobacter pylori a cause of duodenal ulcer? A methodologic critique of current evidence. Am J Med. 1991;91:566-72.

2. NIH Consensus Development Panel on Helicobacter pylori in Peptic Ulcer Disease. NIH Consensus Conference. Helicobacter pylori in peptic ulcer disease. JAMA. 1994;272:65-9.