Abstract
It has become evident that at least some strains of Helicobacter pylori are pathogenic for humans. However, H. pylori are highly diverse and at least part of this variation involves characteristics related to pathogenicity. A large amount of evidence suggests that H. pylori infection of humans is ancient, and in general, the interaction is not terribly destructive. However, we should try to understand better the risks associated with infection with particular H. pylori strains, and limit treatment to those situations in which the indications for eradicating H. pylori are clear-cut. It is entirely possible that some H. pylori strains are commensals, and that others are symbionts. Eradicating those infections could ultimately cause more harm than good. It is too early to reach firm conclusions about whether all H. pylori infections need to be eradicated.