Disease-specific Helicobacter pylori virulence factors: the unfulfilled promise.

A number of putative virulence factors for Helicobacter pylori have been identified including cagA, vacA and iceA. The criteria for a true virulence factor includes meeting the tests of biologically plausibility with the associations being both experimentally and epidemiologically consistent. Although disease-specific associations have been hypothesized/claimed, there are now sufficient data to conclusively state that none of these putative virulence factors have disease specificity. CagA has been claimed to be associated with increased mucosal IL-8 and inflammation, increased density of H. pylori in the antrum, duodenal ulcer (DU), gastric cancer, and protection against Barrett's cancer. Only the increase in IL-8/inflammation is direct and substantiated. Different H. pylori strains with functional cag pathogenicity islands do not vary in virulance as it has been shown that mucosal IL-8 levels are proportional to the number of cagA + H. pylori independent of the disease from which the H. pylori were obtained. It is now known that the density of either cagA + and cagA-H. pylori in the antrum of patients with H. pylori gastritis is the same. In contrast, the mean density of H. pylori in the antrum in DU is greater than in the antrum of patients with H. pylori gastritis. Of interest, the density of H. pylori is higher in the corpus of patients with H. pylori gastritis than those with DU, suggesting that acid secretion plays a critical role in these phenomena. The presence of a functional cag pathogenicity island increases inflammation and it is likely that any factor that results in an increase in inflammation also increases the risk of a symptomatic outcome. Nevertheless, the presence of a functional cag pathogenicity island has no predictive value for the presence, or the future development of a clinically significant outcome. The hypothesis that iceA has disease specificity has not been confirmed and there is currently no known biological or epidemiological evidence for a role for iceA as a virulence factor in H. pylori-related disease. The claim that vacA genotyping might prove clinically useful, e.g. to predict presentation such as duodenal ulcer, has been proven wrong. Analysis of the worldwide data show that vacA genotype s1 is actually a surrogate for the cag pathogenicity island. There is now evidence to suggest that virulence is a host-dependent factor. The pattern of gastritis has withstood the test of time for its relation to different H. pylori-related diseases (e.g. antral predominant gastritis with duodenal ulcer disease). The primary factors responsible for the different patterns of gastritis in response to an H. pylori infection are environmental (e.g. diet), with the H. pylori strain playing a lesser role. Future studies should work to eliminate potential bias before claiming disease associations. Controls must exclude regional or geographic associations related to the common strain circulation and not to the outcome. The authors must also control for both the presence of the factor and for the disease association. The study should be sufficiently large and employ different diseases and ethnic groups for the results to be robust. The findings in the initial sample (data derived hypothesis) should be tested in a new group (hypothesis testing), preferably from another area, before making claims. Finally, it is important to ask whether the results are actually a surrogate for another marker (e.g. vacA s1 for cagA) masquerading for a new finding. Only the cag pathogenicity island has passed the tests of biological plausibility (increased inflammation) and experimental and epidemiological consistency.