Scandinavian Journal of Gastroenterology 24(5): 599-606, 1989.
Kronborg O, Fenger C, Olsen J, et al.
Two screenings with Hemoccult-II were performed in a randomized study, allocating 30,970 persons to screening and 30,968 as controls in a population of 140,000 between 45 and 74 years old on the island of Funen, Denmark. The test was completed in 20,672 initially and in 18,779 of these during rescreening 2 years later. Positive H-II tests were found in 215 and 159 persons during the two screenings, respectively. A total colonoscopy was performed in 187 and 144, and cancer was detected in 37 and 13 and adenomata in 86 and 76, respectively. Interval cancers had developed in 40 persons at the end of the second screening, and 39 non-responders had developed cancer. Cancer was diagnosed in 115 controls and an adenoma in 100 during the same period. Interval cancers presented as rectal cancers more frequently than those detected by screening. Early cancers were more frequent in the screening group, and, accordingly, more patients had curative and also less extensive surgery, with a low postoperative mortality. The total number of deaths from colorectal cancer was 37 in the screening group, including interval cancers and cancers in non-responders and persons who developed cancer before they could be invited, which suggests a reduction in mortality of 27% (51 deaths among controls, compared with 37). The reduction is as yet not statistically significant, and final evaluation must await at least one more screening, ending in 1990, and a follow-up of some years. Removal of more large adenomas during screening makes it possible that the incidence of cancer will decrease.
Rheinische Friedrich- Wilhelms- Universität Bonn