The infection by Helicobacter pylori arises a question of great interest: should the infection be considered a disease in all cases? or, on the contrary, is the microorganism harmful in only some circumstances? The response to these questions is followed by therapeutic implications of great importance, as it is followed by the decision of administering eradication therapy to every patient or only to some individuals. In this article the arguments supporting each of the aforementioned options are reviewed. Firstly, the aphorism recently stated <> suggests that all microorganisms are harmful and, therefore, generalized H. pylori eradication therapy should be administered to all infected subjects. Among the arguments supporting this attitude are the following: the H. pylori <> carrier status does not exist; a high risk of developing peptic ulcer exists in infected patients; H. pylori positive patients are at risk of suffering gastric adenocarcinoma and lymphoma; and, finally, the eradication of the microorganism would reduce the potential reservoir of the infection. On the other hand, the expression <> summarises a more conservative position, that suggests the existence of <>, <> H. pylori, advising the eradication of the microorganism in only some cases. Several arguments supporting this idea, and, therefore, against the generalized use of eradication therapy, have been proposed: H. pylori has colonized humans since immemorial time and, therefore, it can not be very harmful to the human; the bacterium will induce peptic ulcer or gastric cancer in only a minority of infected subjects; the beneficial effect of eradication has been definitively demonstrated in few entities as gastroduodenal ulcer and MALT lymphoma; some strains could be beneficial to the human; generalized administration of antibiotic therapy will cause problems; the H. pylori vaccine represents a more rational alternative; and, finally, the incidence of H. pylori infection is spontaneously decreasing in some regions. In summary, there are strong arguments supporting both of the strategies that are reviewed in the present article, and, therefore, the dilemma <> remains.
