Background: Knowledge on the genetic risk of peptic ulcer has predominantly been based on hospital materials. To minimize selection bias, we tested the association between some genetic markers and the risk of peptic ulcer in a large-scale epidemiologic design., Methods: Some 3387 white men aged 55-74 years were investigated and reported their history of peptic ulcer. Information about hospitalization and operation was collected from registers., Results: The lifetime prevalence of peptic ulcer in men with the Lewis phenotype Le(a + b-) and non-secretors of ABH antigen was 15%, significantly higher than others, 11% (P < 0.01); the risk in phenotypes O and A were equally high, 12%, and among other ABO groups, 7% (P < 0.05). Men with phenotype O had significantly higher risk of hospitalization than others (P < 0.01). Compared with others, the attributable risk of peptic ulcer in men who were Le(a + b-) or non-secretors, with O or A phenotypes, was 37%. No association was found with complement C3, MNS, or Rhesus blood groups., Conclusions: 1) The Le(a + b-) phenotype and the ABH non-secretor trait are relevant genetic markers of peptic ulcer. We suggest that these men have increased susceptibility to Helicobacter pylori infection. 2) This study challenges the importance of the ABO blood group: lifetime prevalence was equally high among men with O and A phenotypes, with more severe cases in men with phenotype O.