Screening for colorectal carcinoma: an analysis of the sensitivity of haemoccult.

Abstract

The sensitivity of Haemoccult for asymptomatic colorectal carcinoma has been estimated in a large randomized controlled trial of mass population screening, with a minimum follow-up of 2 years. A total of 111 cancers were diagnosed in those who completed the screening tests; of these, 36 appear to have been missed by Haemoccult and 75 were detected by the test, giving an overall sensitivity of 67.6 per cent. Haemoccult was shown to be significantly more sensitive for carcinoma of the sigmoid and descending colon than for rectal or right-sided cancers (81 versus 45 and 47 per cent, respectively). The sensitivity was higher when tests were completed over 6 rather than 3 days (74 versus 65 per cent), but this difference was not statistically significant. There was no evidence for a detrimental effect on tumour stage of a false-negative Haemoccult test; indeed, a higher proportion of the interval cancers were Dukes' A tumours than cancers in the control group.