Abstract

The results of screening individuals referred to the Family Cancer Clinic at St Mark's Hospital from 1986 are presented. Colonoscopy was performed in 644 asymptomatic individuals (from 436 families) with a family history of colorectal cancer. Sixty nine (15.8%) of the families fulfilled the Amsterdam criteria for the hereditary non-polyposis colorectal cancer syndromes (HNPCC). Seven cases of colorectal cancer were diagnosed at an average age of 49 years; six at Dukes's stage A and one at stage C, four in subjects from Amsterdam criteria families. One hundred and forty four (22.4%) subjects had one or more adenomas. The prevalence of adenomas in the subjects from Amsterdam criteria families was 34 of 127 (26.8%) compared with 110 of 517 (21.3%) in those from other families; the age and sex adjusted odds ratio (OR) was 1.76 (p = 0.02). Factors influencing the prevalence of adenomas in screened individuals were evaluated. Multivariate analysis showed that independent variables significantly related to the risk of adenomas were: age (p < 0.0001), sex (p = 0.0002), and the number of generations (> or = 2 v 1) of relatives affected by either colorectal cancer or adenomas (p = 0.0006). The latter variable was more highly predictive of the probability of finding an adenoma at colonoscopy than a family history of two generations with cancer only (p = 0.056). The OR of having colorectal adenomas increased with age, by about twofold for each decade, and was twice as high in men than women, and in subjects with two or more generations relative to those with one generation affected by colorectal cancer or adenomas. Six of seven patients with cancer and 46 of 144 (31.9%) with adenomas had lesions proximal to the splenic flexure only. The proportion of individuals with proximal adenomas only was 47.1% in Amsterdam criteria families and 27.3% in the others (p=0.03). These findings support the view that colonoscopy rather than sigmoidoscopy is the method of choice for screening high risk groups.