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


Helicobacter pylori infection increased the risk for gastric carcinoma

ACP J Club. 1992 Jan-Feb;116:27. doi:10.7326/ACPJC-1992-116-1-027

Related Content in this Issue
• Companion Abstract and Commentary: Helicobacter pylori infection increased the risk for gastric carcinoma among Japanese American men

Source Citation

Parsonnet J, Friedman GD, Vandersteen DP, et al. Helicobacter pylori infection and the risk of gastric carcinoma. N Engl J Med. 1991 Oct 17;325:1127-31.



To evaluate the association between gastric carcinoma and previous H. pylori infection in adults.


Case-control study nested within a cohort followed since the 1960s.


Kaiser Permanente Medical Care Program multiphasic health checkups.


128 992 adult subscribers to this program had serum samples frozen and stored between 1964 and 1969. By October 1989, 246 participants had been diagnosed with gastric carcinoma. 200 of these were randomly chosen for investigation and were matched by age at serum sampling, sex, race, and site and date of serum donation to a control from the cohort. 14 pairs had serum samples missing for 1 or both members; thus, 186 pairs (93%) were included in the analysis.

Assessment of risk factors

Serum samples were tested by ELISA for H. pylori IgG. The threshold for a positive ELISA was twice the mean value for 15 known negative controls and was calibrated by these controls and 1 positive control for each run. Assays were done in duplicate, blinded to case-control status. Information on smoking, ulcer disease, gastric surgery, and blood group was obtained from program enrollment questionnaires and health checkups.

Main outcome measure

Gastric carcinoma. There was no pathologic confirmation for 24 patients (13%), and 15 patients (8%) had other or undetermined cancers. The diagnosis of cancer was made 1 to 24 years after serum collection (mean, 14.2 years).

Main results

149 of 186 patients (80%) had serum samples infected with H. pylori compared with 111 controls (60%, matched odds ratio [OR] 2.7, 95% CI 1.6 to 4.3). 84% of 109 patients with confirmed gastric adenocarcinoma had been infected compared with 61% of controls (OR 3.6, CI 1.8 to 7.3). ORs for intestinal and diffuse types of adenocarcinoma were 3.1 (CI 1.5 to 6.6) and 8.0 (CI 1.0 to 64.0), respectively. 91% of 33 women with adenocarcinomas were infected with H. pylori compared with 39% of controls (OR 18.0, CI 2.4 to 134.8). In conditional logistic regression (excluding 7 pairs with nongastric cancers), H. pylori infection (OR 2.7, P < 0.001) and gastric surgery (7 patients; OR 16.8, P = 0.03) were independently associated with cancer, whereas a history of ulcer disease was a protective factor (8 patients and 14 controls; OR 0.2, P = 0.02).


A history of infection with Helicobacter pylori increased the risk for gastric carcinoma.

Sources of funding: Stanford University and the National Cancer Institute.

Address for article reprint: Dr. J. Parsonnet, HRP Building, Room 109A, Stanford University School of Medicine, Stanford, CA 94305-5425, USA.


These 2 case-control studies demonstrate an increased rate of previous exposure to H. pylori among persons with gastric cancer. Unresolved, however, is whether H. pylori has a causal role in gastric cancer. Because H. pylori is so common (occurring in up to 50% of older persons in the United States and in even higher proportions in other countries), it is clearly not sufficient to cause cancer—other features must contribute. Approximately 75% and 60% of the controls were positive for H. pylori in the studies by Nomura and colleagues and Parsonnet and colleagues, respectively. At most, then, H. pylori may be a contributing cause of gastric cancer. Such a relation would be clinically important by suggesting that eradication of H. pylori might prevent cancer. A similar relation of "necessary but not sufficient" may also explain H. pylori's role in peptic ulcer disease (1). However, as Parsonnet and colleagues explain, it is an alternative possibility that " H. pylori might just be a marker [and not a cause] for an increased risk of cancer" if a yet-unidentified factor increases the susceptibility to cancer and to colonization with H. pylori.

The clinical implications of these studies are not clear as a causal relationship is not established. Thus, we do not know whether eradication of H. pylori would reduce cancer risk. Furthermore, any effort to eradicate H. pylori would necessarily involve an enormous number of persons who would need to be treated and re-treated over many years.

Although neither of these studies was designed to permit conclusions regarding the benefits of eradicating H. pylori, some can be inferred. If H. pylori eradication were possible in selected individuals and this resulted in the prevention of gastric carcinoma, the data from Nomura and coworkers provide an optimistic estimate that one would need to treat at least 44 patients to prevent 1 cancer. However, this does not take into account the resources required to identify infected individuals or how long and how often these 44 individuals (who will not develop cancer) will have to be treated over a 20-year period.

In summary, then, these studies provide provocative evidence that H. pylori does precede gastric cancer; duodenal ulcer disease may similarly be preceded by H. pylori. However, the evidence does not yet support a causal role. At most, H. pylori may be "necessary but not sufficient" as a cause. Alternatively, the association may simply be coincidental. We shall watch with interest as the H. pylori story unfolds over the next several years.

David F. Ransohoff, MD
University of North Carolina at Chapel Hill School of MedicineChapel Hill, North Carolina, USA

David F. Ransohoff, MD
University of North Carolina at Chapel Hill School of Medicine
Chapel Hill, North Carolina, USA


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