We appreciate Dr. Pilote's concerns. First, the patients' disease status (duodenal or gastric ulcer) at entry in the cohort was known. As stated in the Methods section of our article, 481 patients (of a cohort of 5924 men) with previous gastrectomy or a previous diagnosis of peptic ulcer disease were excluded from the study. It is standard practice in a cohort study to exclude persons who, at baseline, have the disease being investigated. Consequently, study participants who were known to have peptic ulcer from their interview or from review of hospital records were not included in our study.

Second, it is not clear what Dr. Pilote means by the statement, "... the authors excluded the controls who developed peptic ulcer disease." From the remaining 5443 men in the study, 229 were hospitalized with peptic ulcer disease between 1968 and 1989. These 229 patients were the case-patients for this study, and 229 controls (matched by age at examination and by date of serum collection) were identified among the men who were not hospitalized with peptic ulcer disease. As a result, any patient hospitalized with peptic ulcer during the 21-year surveillance period was identified as a case-patient and was removed from the list of potential controls in the study.

We pointed out in our Discussion that we identified hospitalized patients with peptic ulcer, but, because of practical considerations, we could not identify nonhospitalized patients in whom peptic ulcer disease was diagnosed during the surveillance period. If nonhospitalized patients with peptic ulcer are included in the control group and if a positive association with H. pylori infection also exists among these patients with less complicated disease, the findings in our study underestimate the strength of the association between peptic ulcer disease and previous infection with H. pylori.