Screening for colorectal cancer by means of unhydrated Hemoccult (HO) is in progress in the Province of Florence since 1982. In 1990 rehydrated HO was introduced in the town of Empoli. Five adjacent municipalities where screening had started in 1987 were selected for comparison. In both areas subjects aged 40-70 were invited by mail to undergo the screening protocol. HO-positive subjects were invited to undergo either pancolonoscopy or a combination of left colonoscopy and double contrast barium enema. HO-negative subjects were invited to repeat screening 2 years later. The positivity rate of HO was significantly higher (P < 0.001) for rehydrated (5%) as compared to unhydrated (3.1%) HO. The positive predictive values for cancer (unhydrated: 5.8%; rehydrated: 8.9%) and for adenomas (unhydrated: 26.7%; rehydrated: 25.5%) did not significantly differ. The detection rates of rehydrated HO were significantly higher as compared to unhydrated HO both for cancer (0.37% vs 0.15%; P < 0.05) and adenomas (1.06% vs 0.72%; P < 0.05%). In the present experience rehydration doesn't produce any decrease in the positive predictive value for cancer or adenomas and the increase in the positivity rate appears quite acceptable when considering the significant increase in the detection rates of cancer and adenomas. We conclude that rehydrated HO should be introduced as the standard test for screening in order to increase sensitivity for colorectal cancer and adenomas.