Objective. Tooth loss has been associated with upper gastrointestinal cancer in several studies, but only one previous study used prospectively collected data. The importance of confounding by Helicobacter pylori has not previously been addressed. The objective was to determine the association between tooth loss and upper gastrointestinal cancer in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study cohort and to determine the importance of potentially confounding dietary factors or H. pylori seropositivity. Material and methods. A prospective cohort study with 29,124 subjects included 49 esophageal squamous cell carcinomas, 66 esophageal/gastric cardia adenocarcinomas, and 179 gastric non-cardia adenocarcinomas occurring between 1985 and 1999. Cox proportional hazards models adjusted for age and education were used to estimate hazard ratios (HRs) and 95%CIs. Odds ratios and 95%CIs were calculated with and without adjustment for H. pylori seropositivity in a nested case-control group to determine whether H. pylori confounded the association between tooth loss and gastric cancer. Results. Tooth loss significantly increased the hazard ratio for gastric non-cardia cancer, the HR (95%CI) for edentulous subjects versus those with < 10 teeth lost was 1.65 (1.09, 2.49, respectively). No statistically significant associations were found between tooth loss and esophageal squamous cell carcinoma or esophageal/gastric cardia adenocarcinoma. Confounding by dietary factors, tobacco smoking, or H. pylori did not explain these results. Conclusions. Tooth loss was associated with increased risk of gastric non-cardia cancer, but not esophageal squamous cell carcinoma or esophageal/gastric cardia adenocarcinoma in this Finnish cohort. Objective importance of confounding by Helicobacter pylori has not been addressed. The objective was to determine the association between tooth loss and upper gastrointestinal cancer in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study cohort and to determine the importance of potentially confounding dietary factors or H. pylori seropositivity. Material and methods. A prospective cohort study with 29,124 subjects included 49 esophageal squamous cell carcinomas, 66 esophageal / gastric cardia adenocarcinomas, and 179 gastric non-cardia adenocarcinomas between 1985 and 1999. Cox proportional hazards models adjusted for age and education were used to estimate hazard ratios (HRs) and 95% CIs. Odds ratios and 95% CIs were calculated with and without adjustment for H. pylori seropositivity in a nested case-control Results from Tooth loss significantly increased the hazard ratio for gastric non-cardia cancer, the HR (95% CI) for edentulous subjects versus those with <10 teeth lost was 1.65 (1.09, 2.49, respectively). No significant associations were found between tooth loss and esophageal squamous cell carcinoma or esophageal / gastric cardia adenocarcinoma. Confounding by dietary factors, tobacco smoking, or H. pylori did not explain these results. Conclusions Tooth loss was associated with increased risk of gastric non-cardia cancer, but not esophageal squamous cell carcinoma or esophageal / gastric cardia adenocarcinoma in this Finnish cohort.