OBJECTIVE To compare two immunochemical faecal occult blood tests based on reversed passive haemagglutination (RPHA) or latex agglutination (Hdia) in a population based screening setting. METHODS Hdia was interpreted according to three positivity thresholds: 100, 150, or 200 ng of haemoglobin/mg of specimen solution. A total of 5844 subjects were recruited into the study, from 17432 invited subjects aged 50-70. RESULTS Positivity rates were 3.3% for RPHA, Hdia100 3.5%, Hdia150 2.5%, Hdia200 2.0%. Among subjects complying with the diagnostic work up, colorectal cancer (CRC) was detected in 19 subjects (17 RPHA positive, 16 Hdia100 positive, 15 Hdia150 positive, 14 Hdia200 positive) and high risk adenoma/s in 41 subjects (28 RPHA positive, 32 Hdia100 positive, 29 Hdia150 positive, 25 Hdia200 positive). The prevalence of screen positive CRC in the population was for RPHA 2.9 per thousand, Hdia100 2.7 per thousand, Hdia150 2.6 per thousand, Hdia200 2.4 per thousand. The prevalence of screen positive high risk adenomas in the population was for RPHA 4.8 per thousand, Hdia100 5.5 per thousand, Hdia150 5.0 per thousand, Hdia200 4.3 per thousand. CONCLUSIONS Hdia100 was as sensitive as RPHA for cancer and high risk adenomas. As Hdia is less technically complex than RPHA, it is a valid alternative to the latter, provided that full automation of the development procedure is available. Increasing the positivity threshold of Hdia up to 150 or 200 ng of haemoglobin/mg of specimen solution is not advisable as the increase in specificity is too small to justify the corresponding decrease in the detection of screen positive cancers in the population.