The aim of our case-control study was to estimate the effect on mortality from colorectal cancer (CRC) of a population-based screening with a faecal occult blood test started in 1982 in a rural area of the district of Florence. We examined the relationship between mortality and the interval since the most recent screening. The cases in the study were 206 individuals who had died from CRC after the age of 41 years. Five controls were selected randomly from the list of individuals alive at the time of diagnosis of the corresponding case and were matched by sex, age and place and length of residence. After adjustment for potentially confounding factors, the odds ratio (OR) for death from CRC for screened persons vs. those not screened was 0.60 [95% confidence interval (CI), 0.4-0.9]. The OR was lowest in the first 3 years after the most recent test (OR = 0.54; 95% CI, 0.3-0.9) and increased towards unity subsequently. Our results suggest that screening for CRC by biennial faecal occult blood testing can reduce mortality from the disease.