Background : Screening with Hemoccult-II ® (H-II) followed by colonoscopy, when fecal occult blood is demonstrated, reduces mortality from colorectal cancer. Whether upper gastrointestinal investigation is necessary when colonoscopy does not reveal any significant colorectal lesion is doubtful, and is the subject of this study. Material : In 1985, 30,967 persons from the general population register of Funen were randomized to biennial H-II screening. A positive test was followed by colonoscopy and no attempt was made to evaluate the upper gastrointestinal tract. Based on the information from the Funen Patient Database, the National Board of Health's Register of Death Causes, the Cancer Register and the National Register of Patients, all persons with malignancy of the gastrointestinal tract were identified. Results : During 15 years and 8 screening rounds, 1,767 tests were positive; 1,536 complete colonic investigations detected colorectal cancer in 182 persons, adenoma ≥10 mm in 440 persons, and in 879 investigations no colorectal lesion was found. Upper GI cancers were diagnosed in 209 persons within 2 years of the H-II test (199 after a negative H-II and no more than 10 persons within 2 years of a positive test). Among the 10, two were diagnosed as a consequence of symptoms at the time of screening. Conclusion : It is unjustified to perform upper gastrointestinal investigation in asymptomatic persons with a positive H-II in a Danish population screening for colorectal cancer.