Helicobacter pylori infection is on the decline in the Western world.1 This is most likely related to improvements in sanitary conditions and socioeconomic status along with the widespread use of antibiotics. Hospitalisation and mortality rates for duodenal ulcer disease, gastric ulcer disease, and cancer of the gastric antrum and corpus, all clearlyH pylori related diseases, also have declined markedly between 1970 and 1995.2 In sharp contrast, hospitalisation and mortality rates for both gastro-oesophageal reflux disease (GORD) and oesophageal adenocarcinoma have increased during the same time interval. Although the hypothesis seemed quite radical in 1997,3 it appears that the decline in H pylori infection may explain these two opposing time trends.

Although initial studies were inconsistent,4 the majority of recent evidence from the USA,5 ,6 Europe,7and the Far East8 find that H pylori infection alone, regardless of virulence factors, protects from the development of severe reflux oesophagitis and its complications, such as Barrett's oesophagus and adenocarcinoma. Even more striking and consistent has been the protection reported from GORD related complications in patients with H pylori infection and the cagA⁺ virulent factor. Vicari and colleagues9 from the Cleveland Clinic were the first to report that the prevalence of the more virulent cagA⁺ strain of H pyloridecreased in patients with more severe complications of GORD. They found that the likelihood of having Barrett's oesophagus complicated by dysplasia or cancer was reduced more than twofold in patients infected with the cagA⁺ strain compared with other groups (odds ratio 0.43, 95% confidence interval (CI) 0.31–0.60). Around the same time, Chow and colleagues10 reported that carriage of the cagA⁺ strains was associated with an increased risk for distal gastric cancer but with a reduced risk for oesophageal and gastric cardia adenocarcinoma. Subsequently, other large studies from the Netherlands11 and in this issue ofGut the UK,12 confirmed the protective role of the cagA⁺ virulent factor (see page341). Additionally, this last study found that women were less likely to suffer from oesophagitis then men, while male sex, age over 50 years, and the presence of a hiatal hernia increased the risk of reflux oesophagitis.

The pathophysiological mechanisms by which H pylori infection protects patients from developing severe GORD have been more slow to be defined. Initial reports from Vicarri and colleagues9 and El Serag and colleagues13suggested that the presence of corpus gastritis, often associated with atrophy and intestinal metaplasia, was protective. In this issue ofGut,14 a large population study of nearly 6000 Japanese subjects found a significant negative correlation between serum pepsinogen level and reflux oesophagitis (odds ratio 0.35, 95% CI 0.18–0.68) (see page 335). Serum pepsinogen I ⩽70 ng/ml and pepsinogen I/II ratio <3.0 had a sensitivity of 70% and specificity of 97% for atrophic gastritis compared with histology in Japan.15

These studies implied that the final denominator of the protection against GORD was decreased gastric acid secretion but the critical link was not established until the report from Japan by Koike and colleagues,16 also in this issue ofGut (see page 330). As formal gastric analysis studies are difficult to perform in large population studies, the authors developed an endoscopic gastrin test. After an overnight fast, subjects are injected intramuscularly with tetragastrin and gastric fluids collected endoscopically 20 and 30 minutes later and analysed for volume and hydrogen ion concentration. The endoscopic gastrin test values correlate very closely with peak acid output determined by conventional methods.17 Using this test, Koike et al found that erosive reflux oesophagitis occurred most often in the absence ofH pylori infection and gastric hyposecretion. Even in the presence of H pylori infection, reflux oesophagitis was most likely to develop in the patient with less severe gastritis and atrophy of the corpus as well as when gastric acid secretion was higher.16

The combination of these studies confirm that the distribution and severity of H pylori related gastritis and atrophy rather than the mere presence of H pylori infection plays a critical role in the pathophysiology of GORD.18 Antral predominant gastritis is probably the most important predictor of duodenal ulcer disease in the presence ofH pylori infection. Antral predominant gastritis leads to reduced antral D cell density and somatostatin concentrations. This subsequently leads to higher serum gastrin levels, which in the presence of a healthy corpus and parietal cell mass, results in increased gastric acid secretion and volume, leading to a higher duodenal acid load and duodenal ulcerations. In contrast, more severe corpus gastritis is associated with lower acid output. Compared with cagA⁻ strains, cagA⁺ strains are associated with enhanced development of atrophic gastritis wherever the organism is present in the stomach. Atrophy of the corpus leads to the destruction of gastric glands and later hypochlorhydria. The end result is that the acid load presented to the duodenum and oesophagus is markedly diminished, thereby being potentially protective for the development of GORD and its complications in subjects with lower oesophageal sphincter dysfunction or hiatal hernia. However, the inflammatory changes in the corpus return to normal when the infection is cured, increasing acid secretion and possibly contributing to the recent reports of oesophagitis developing in healthy subjects and duodenal ulcer patients after successful treatment ofH pylori infection.19 ,20

These results confirm a protective role of H pylori for reflux oesophagitis, and are consistent with epidemiological data from Western countries showing an increase in GORD and decrease in gastric cancer in concert with the decline inH pylori infection. Therefore, there appears to be both risks as well as benefits to the indiscriminate worldwide attempt to eliminate this organism. As such, the concept that the only good H pylori is a deadH pylori needs to be revisited.21

Perhaps no one can summarise the complexity of H pylori infection better than Martin Blaser who said “ . . .. those looking for simple answers about the relations of H pylori and disease undoubtedly will be disappointed; the complexity likely is older than the human race”.22