Article Text


Screening for early gastric cancer
  1. K E L McColl
  1. Correspondence to:
    Professor K E L McColl
    Section of Medicine, Western Infirmary, 44 Church St, Glasgow, UK; k.e.l.mccoll{at}

Statistics from

Can serological testing for Helicobacter pylori infection and atrophic gastritis predict the risk of gastric cancer?

Gastric cancer remains one of the most common malignancies in the world.1 The great majority of cases are fatal because the tumour has usually reached an incurable stage by the time of diagnosis.2 The early potentially curable stage of the disease is usually asymptomatic and consequently patients mainly present when they have developed symptoms of advanced or complicated disease.2,3 There is considerable interest in a means of detecting this common cancer at an early and curable stage.

In the Western world, the incidence of gastric cancer distal to the cardia has fallen over recent decades whereas that at the cardia and gastro-oesophageal junction has increased markedly.4 However, at a global level, non-cardia gastric cancer remains the predominant type and it is this which we are discussing.1

Major advances have been made in our understanding of the aetiology and pathogenesis of non-cardia gastric cancer and of the precancerous changes which occur in the gastric mucosa. Helicobacter pylori infection is now accepted as an essential cofactor in the majority of cases of non-cardia gastric cancer.5 However, detecting the presence of the infection is, in itself, of limited value in predicting cancer occurrence; more than 50% of the world’s population have the infection and the lifetime risk of an infected individual developing the cancer is less than 2%.

Recent studies have indicated that the risk of a person with H pylori infection developing non-cardia gastric cancer is highly dependent on the pattern of gastritis and altered physiology which the infection induces in their stomach.6 In some patients, the infection induces an antral predominant non-atrophic gastritis associated with normal or high acid secretion. Patients with this gastric phenotype may develop duodenal ulcers but very rarely develop gastric cancer. Other patients develop an atrophic pangastritis and reduced gastric acid secretion and it is this gastric phenotype which is associated with gastric cancer. These two phenotypes represent ends of a spectrum, with many subjects showing an intermediate disease pattern. It is also possible for patients to progress from the non-atrophic to atrophic pattern with advancing years.7,8

In a large prospective study in Japan, Uemura et al endoscopically monitored patients with H pylori infection and different gastric phenotypes for a mean of 7.8 years.6 None of the duodenal ulcer patients developed gastric cancer. However, 4.7% of non-ulcer dyspepsia patients developed cancer and its occurrence was strongly associated with the presence of atrophic pangastritis detected at the initial endoscopy. Such observations have stimulated interest in using the H pylori induced gastric phenotype to direct screening for early gastric cancer.

H pylori infection can be diagnosed non-endoscopically by means of serology. It is also possible to detect atrophic gastritis serologically. The concentration of pepsinogen I and the ratio of pepsinogen I/II are low in patients with atrophic gastritis.9,10 Consequently, simple blood tests can determine whether a subject has H pylori infection and whether the infection has induced the atrophic pattern of gastritis which is associated with a high risk of gastric cancer.

In this issue of Gut, Watabe and colleagues11 present a large prospective study in which they have assessed the value of combined serological testing for H pylori infection and atrophic gastritis in predicting the risk of gastric cancer (see page 764). The study included almost 10 000 members of the general population who had annual endoscopic examinations for a mean of five years following serological tests. They analysed the outcome by dividing subjects into four groups according to whether they were H pylori seropositive or negative and whether they had a low pepsinogen I and I/II ratio, indicating atrophic gastritis.

The risk of developing gastric cancer in H pylori negative patients without atrophic gastritis was similar to that in H pylori positive patients without atrophic gastritis. Relative to these two groups, the risk of cancer was sixfold higher in H pylori positive patients with atrophic gastritis and eightfold higher in H pylori negative patients with atrophic gastritis. Age and male sex were also independent risk factors for cancer occurrence.

This study provides useful confirmation of the importance of the gastric phenotype induced by H pylori infection in determining the risk of gastric cancer. Patients with H pylori infection but no evidence of atrophy had a similar cancer risk to those with an uninfected healthy stomach, at least over the subsequent five years examined in this study. Patients with atrophic gastritis had a markedly increased cancer risk irrespective of H pylori serology. Atrophy in those with negative H pylori serology can be largely explained by the tendency for the infection to disappear in the atrophic achlorhydric stomach.12 Due to this phenomenon, one can conclude that H pylori serology on its own is of limited value in determining the risk of imminent development of gastric cancer and that detecting atrophic gastritis by serum pepsinogens is the important test.

How useful would such serological testing be in identifying the subgroup of the population most likely to develop cancer within the subsequent five years and who might benefit from careful surveillance endoscopy? The data presented by Watabe and colleagues11 indicate that serological screening for low pepsinogen I would allow identification of a subgroup of 22% of the population who would go on to develop 70% of the gastric cancers occurring within the following five years. The annual incidence of cancer in subjects in the high risk groups in this Japanese based study was approximately 0.5% and thus equivalent to the risk of cancer occurrence in Barrett’s oesophagus patients studied in the West.13

Identifying subjects at high and imminent risk of developing gastric cancer is only useful if one can intervene to improve the natural history of the disease. Such intervention at present involves annual endoscopies to detect and treat the tumours at an early stage of their development. In this study reported by Watabe and colleagues,11 such a strategy allowed them to detect all of the cancers which developed in the high risk groups while still localised within the submucosa. Approximately 50% of the tumours were treated by endoscopic resection and the remainder underwent surgical operation. The ability to detect early tumours endoscopically depends on adequate endoscopic facilities and operator skills. Although these are undoubtedly common place in Japan, they may not be so readily available in other countries with a persistent high incidence of gastric cancer. However, one great attraction of the pepsinogen serological screening is that it allows the available resources to be concentrated on the group most likely to benefit from them and so increases cost effectiveness.

Although the ability to detect gastric cancer at an early curable stage is attractive, it is not always in the individual’s best interests. The natural history of early gastric cancer means that left untreated it usually does not cause clinical problems for several years and indeed may not cause any symptoms during the patient’s natural lifespan.14,15 Surgical gastric resection is associated with morbidity and a small mortality. Although endoscopic resection is likely to be associated with less morbidity and mortality, the efficacy of this less radical management in achieving long term cure is less clear. Screening and surveillance for early gastric cancers would be more attractive if we had a simple, safe, and effective way of managing them.

The ideal way to reduce the mortality due to gastric cancer is to interrupt the precancerous process before an actual cancer develops. Prevention is better than cure. Current understanding is that H pylori infection is usually an essential cofactor in initiating the sequence of events that finally lead to the cancer.16 One would therefore anticipate that preventing people contracting the infection in the first place or treating the infection at an early stage before irreversible atrophic gastritis develops would be the most effective and cost effective way of reducing the incidence of gastric cancer. Japan and other countries with a persistent high incidence of non-cardia cancer would be particularly appropriate for such a prophylactic interventional strategy.

However, even if every H pylori infected subject in the world had their infection eradicated today, the incidence of gastric cancer would probably remain high for several decades to come. This is due to the fact that eradicating the infection does not produce resolution of atrophic gastritis and subjects with these irreversible changes will continue to be at risk of gastric cancer. Consequently, there will continue to be a need for more efficient ways of identifying subjects at high and imminent risk of developing this sinister tumour.

Can serological testing for Helicobacter pylori infection and atrophic gastritis predict the risk of gastric cancer?

View Abstract


  • Conflict of interest: None declared.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.