Background: Several measures of familial disease aggregation have been proposed, but only a few of these are designed to be implemented at the individual level. We evaluate two of them in the context of breast-cancer incidence.
Methods: A population-based cohort consisting of 114 429 women born between 1874 and 1931 and at risk for breast cancer after 1965 was identified by linking the Utah Population Data Base and the Utah Cancer Registry. Two competing methods were used to obtain predictors of familial aggregation of risk: the number of first-degree relatives with breast cancer (NIST) and the familial standardised incidence ratio (FSIR), which weights the disease status of relatives based on their degree of relatedness with the proband. Relative risks were estimated using Mantel-Haenszel. Poisson regression and spline regression methods. The age-dependent hazard function was also estimated.
Results: Compared to a baseline category containing 91.5% of the subjects, the 0.7% of subjects identified as high risk using the FSIR criterion had a relative risk of about 2.8, while those identified as high risk using the NIST criterion had a relative risk of 2.0. Moderate-risk subjects had a relative risk of about 1.75 using either criterion. FSIR was a significant predictor of risk even for those with no affected first-degree relatives. No decline in the baseline risk was observed at advanced ages.
Conclusions: FSIR appears to be a better predictor of breast-cancer risk than NIST, particularly for high-risk subjects.