This study was undertaken to determine whether selected risk factors for esophageal and gastric cancer are associated with tumors that overexpress cyclin D1. Archived tumor tissue was available for 630 esophageal and gastric cancer patients who participated in a population-based case-control study. Patients were categorized into case groups based on whether protein overexpression of the cyclin D1 gene, as assessed by immunohistochemistry, was present (cyclin D1+, n = 285) or not (cyclin D1-, n = 345) in the tumor. The distribution of risk factors in each of these case groups was then compared with the distribution among the 695 controls. Multivariate-adjusted odds ratios (OR) for esophageal adenocarcinoma were reduced in relation to use of aspirin and other nonsteroidal anti-inflammatory drug (NSAID) use but only among patients with cyclin D1+ tumors (0.45, 95% confidence interval [CI] = 0.26, 0.79) and not among those with cyclin D1- tumors (1.12, 95% CI = 0.67, 1.86). A similar pattern was observed for gastric cardia adenocarcinomas. In contrast, ORs for esophageal squamous cell carcinoma and noncardia gastric adenocarcinomas in relation to NSAID use were reduced, regardless of cyclin D1 status. ORs did not vary with cyclin D1 status in relation to alcohol, body size, or cigarette smoking, with the following exception; for noncardia gastric adenocarcinomas the cyclin D1- tumors showed a 2-fold elevation in the OR with ever smoking. These data suggest that the reduction in risk associated with NSAID use may be restricted to those esophageal and gastric cardia adenocarcinomas that overexpress cyclin D1.