Background: An association between pernicious anemia and stomach cancer has been established in several studies. An increased risk of pancreatic and esophageal cancers has also been reported among pernicious anemia patients. The aim of this case-cohort study was to identify additional risk factors for cancer of the esophagus, stomach, and pancreas among patients with pernicious anemia. Methods: A population-based cohort of 4586 patients with pernicious anemia was linked to the Swedish Cancer Registry to identify patients who subsequently developed cancers of the esophagus, stomach, or pancreas using a case-cohort design. A subcohort consisting of 4% of the cohort was randomly selected to serve as the comparison group. Information on medical history, smoking habits, and alcohol use was retrieved from medical charts and analyzed for cancer patients and subcohort members. Results: We could not identify any risk factors other than pernicious anemia for stomach cancer. For pancreatic and esophageal cancer, younger age at diagnosis of pernicious anemia was associated with an increased risk. A prior gastric resection, smoking and alcohol abuse were more frequent among esophageal cancer cases than in the subcohort. Conclusions: We conclude that a causal relationship between pernicious anemia and subsequent development of esophageal or pancreatic cancers still remains unproven. For esophageal cancer, confounding by smoking and alcohol use is the likely explanation of earlier reports of an association. In the case of stomach cancer, both the inflammatory process, secondary to the pernicious anemia, and pernicious anemia per se may be factors leading to malignant transformation.