Screening for occult blood by means of guaiac tests has an unsatisfactory sensitivity for the detection of colorectal neoplasms. To increase sensitivity and specificity the immunological determination of human hemoglobin and albumin in feces has been developed. The validity of analyzing only two samples from one bowel movement of either test is not known. An immunological determination of human fecal hemoglobin and albumin using luminescence immunoassays (LIA) was performed in 739 patients with gastrointestinal complaints before scheduled colonoscopy. Each patient collected two 1 ml samples from one stool. There were no dietary restrictions. The sensitivity for detecting colorectal carcinomas was 95.3% (95% confidence interval 84.2-99.4%) with hemoglobin and 67.4% (95% confidence interval 51.2-80.9%) with albumin. The sensitivity for detecting large adenomatous polyps was 62.9% (95% confidence interval 50.5-74.1%) with hemoglobin and 32.9% (95% confidence interval 22.1-45.1%) with albumin. The specificity was 97% for hemoglobin, 96% for albumin and 94% for the combined test. The immunological determination of fecal hemoglobin is superior to albumin and has a better sensitivity for the detection of colorectal neoplasms than that reported for guaiac tests, even if two samples from one bowel movement are examined. The immunological determination of fecal hemoglobin should therefore be evaluated for use in colorectal cancer screening.