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


Etiology

Helicobacter pylori and duodenal ulcer: a review

ACP J Club. 1992 Mar-April;116:61. doi:10.7326/ACPJC-1992-116-2-061


Source Citation

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


Abstract

Objective

To review the association of Helicobacter pylori with chronic active gastritis and duodenal ulcer and to assess if H. pylori causes duodenal ulcers.

Data sources

A MEDLINE search for 1985 to 1990 and review of The Lancet, The New England Journal of Medicine, Annals of Internal Medicine, Gastroenterology, and Current Contents from 1988 to 1990 identified clinical studies of H. pylori and duodenal ulcers.

Study selection

Published reports of clinical studies with at least 25 patients, comparison groups, data on H. pylori for the groups being compared, and random treatment assignment for therapeutic studies were included. 6 cross-sectional and 2 follow-up studies were evaluated.

Data extraction

Data to support Bradford Hill's principles for assessment of possible cause-and-effect relationships were extracted from each study: strength of association, dose-response relationship, and temporal relationship between the causal agent (H. pylori) and the effect (duodenal ulcer), and whether the removal of the putative agent prevents the effect.

Main results

The cross-sectional studies showed that 78% to 100% of patients with duodenal ulcers had antral H. pylori infection whereas 3% to 70% of patients without duodenal ulcers had H. pylori. 98% of patients with duodenal ulcer also had chronic active gastritis, and 80% to 98% of patients with gastritis without duodenal ulcers had H. pylori. The studies did not grade the severity or extent of gastric H. pylori and could not determine whether infection preceded ulceration. The 2 randomized controlled studies measured ulcer healing and recurrence after therapy for removal of H. pylori. Healing rates (59% to 74%) were similar after treatment with cimetidine or colloidal bismuth subcitrate, with or without tinidazole. The relapse rates for those patients with persistent H. pylori after treatment were 76% and 84% and for those without H. pylori were 10% and 21%. These lower relapse rates may be due to eradication of the organism or to a direct mucosal effect of treatment.

Conclusions

Although the association of H. pylori with duodenal ulcer and chronic active gastritis is strong, the evidence does not support the conclusion that H. pylori causes duodenal ulcers. More research must be done to assess the temporal relationship of gastric H. pylori infection with chronic active gastritis and duodenal ulcers.

Source of funding: In part, Robert Wood Johnson Foundation.

Address for article reprint: Dr. L. Rabeneck, Veterans Affairs Medical Center (111D), 2002 Holcombe Boulevard, Houston, TX 77030.


Commentary

H. pylori, a spiral-shaped bacterium found in the antrum of most patients with inflammatory gastroduodenal lesions, is generally accepted as a cause of chronic active gastritis; its importance in the pathogenesis of duodenal ulcers remains unclear.

To assess the role of H. pylori in the etiology of duodenal ulcers, Rabeneck and Ransohoff use 4 (of 9) principles of Bradford Hill (1) (association, biologic gradient, temporality, and experimental data) in a critical review of 8 clinical studies. The advantage of Hill's approach is to permit evaluation of the evidence in context, ignoring the differences in methods, without the requirement of sine qua non proof from each principle and without reliance on formal tests of significance. This method of analysis of H. pylori and duodenal ulcers shows only association, without the support of causality by temporality or biologic gradient. The authors reject analysis by Koch's postulates because of their "limited applicability to problems in chronic disease." This rejection of Koch's postulates, however, is unwarranted because tuberculosis, the locus classicus of the postulates, is itself a chronic disease: Previous attempts to satisfy Koch's postulates have failed to establish a cause-and-effect relationship for H. pylori and duodenal ulcers (2).

Regardless of the methodologic approach, current evidence does not establish H. pylori as an unequivocal cause of duodenal ulcers. Until more data are available (or more investigators swallow the organism [2]), clinicians should continue to treat patients with conventional anti-ulcer therapy. Antimicrobial therapy for duodenal ulcers should be reserved for clinical trials for selected patients with unusual, refractory ulcer disease.

I. David Shocket, MD
Department of Veterans Affairs Medical Center Boston, Massachusetts