The potential association of polymorphism in the HRAS protooncogene variable repeat region with susceptibility to cancer has become a controversial topic. A number of studies have produced results that appear inconsistent. We report here a multidisciplinary study with a combined molecular and epidemiological approach, addressing the specific question of the association of rare HRAS alleles and breast cancer. Extensive questionnaire data and peripheral blood for DNA extraction were obtained from 160 cases of incident breast cancer and from two control groups totaling 405 unaffected women from five outpatient clinics in North Carolina between April 1990 and June 1991. Controls were frequency matched to cases on age and race. Our results, adjusted for race and age, showed a positive overall association between the presence of rare HRAS alleles and breast cancer. This relationship was somewhat stronger in control group 2 (odds ratio = 3.0; P < 0.01) than in control group 1 (odds ratio = 2.0; P < 0.05). The relationship was 3-6 times stronger in blacks than in whites. In the case series, rare HRAS alleles were associated with hormone receptor negative tumors. This association was stronger in blacks and younger women. There was no confounding or effect modification by any other breast cancer risk factors. We conclude that rare HRAS alleles are associated with breast cancer and that this association may be stronger in black women than in white women. Rare HRAS alleles may also be related to more aggressive tumors, particularly in blacks and younger women. HRAS alleles have the potential to become a valuable screening biomarker for women at increased risk for breast cancer.