Newfoundland has the highest rate of colorectal cancer (CRC) of any Canadian province. In order to investigate the factors, especially genetic components, responsible for CRC we established the Newfoundland Colorectal Cancer Registry. In a 5-year period we examined every case of CRC diagnosed under the age of 75 years and obtained consent from 730 cases. Careful analysis of family history was used to assign a familial cancer risk, based on established criteria. We observed that 3.7% of CRC cases came from families meeting the Amsterdam II criteria and a further 0.9% of cases involved familial adenomatous polyposis (FAP). An additional 43% of cases met one or more of the revised Bethesda criteria and 31% of all cases had a first-degree relative affected with CRC. We compared the Newfoundland data with data from the province of Ontario, where the same recruitment and risk-assessment criteria were used. In all categories, the indicators of familial risk were significantly higher in Newfoundland. These data were also compared to results published from 13 other population-based studies worldwide. In every category the proportion of Newfoundland cases meeting the criteria was higher than in any other population. The mean differences were: 3.5-fold greater for FAP, 2.8-fold higher for Amsterdam criteria, 2.0-fold higher for Bethesda criteria and 1.9-fold higher for the number of affected first-degree relatives. We conclude that the high incidence of CRC in Newfoundland may be attributable to genetic, or at least familial, factors. In the high-risk families we provide evidence for the involvement of founder mutations in the APC and MSH2 genes.

Hereditary non-polyposis colon cancer (HNPCC) is a significant cause of colorectal and other malignancies. Due to the lack of features that reliably differentiate between a sporadic case and an inherited case of colon cancer, it is likely that HNPCC is under reported. The diagnosis of HNPCC relies heavily on finding multiple cases of colorectal or other specific cancers within a family. In the absence of a significant family history, a diagnosis of HNPCC is seldom considered. We postulate that small kinships--or, more specifically, kinships with a low information content--are more likely to be designated as having a low risk of an inherited cancer predisposition than are large kinships. This has the potential to exacerbate the under-diagnosis of HNPCC in small families, leading to inadequate treatment, follow-up and family counselling. We have developed an objective measure of the information content of individual pedigrees called the Sum of Information on Seventy-yr-old Equivalents (SISE) coefficient. The SISE coefficient is a function of the number of relatives in a kinship and their relationship to the proband, of their ages and of the age-dependent penetrance of HNPCC mutations. A population-based series of colorectal cancer cases was assessed, by currently accepted methods, for the likelihood of there being an HNPCC mutation segregating in each family. We observed that families with a low SISE coefficient were significantly more likely to be designated at low risk of HNPCC (P< or =0.001). Using a cumulative binomial distribution function, we estimated the likelihood of observing multiple cancers in families of different SISE coefficients.