Background. Systemic Autoimmune Rheumatic Diseases (SARDs) have been inconsistently associated with breast cancer risk. We evaluated this association in a population of U.S. elderly women. Methods. We designed a population-based case-control study of elderly women using SEER-Medicare linked database. Female incident breast cancer cases, diagnosed between 1993 and 2002 (n=84,778), were compared to an equal number of age-matched cancer-free female controls. History of SARDs including rheumatoid arthritis (RA, n=5,238), systemic lupus erythematosus (SLE, n=340), Sjogren's syndrome (SS, n=374), systemic sclerosis (SSc, n=128) and dermatomyositis (DM, n=31) from age 65 to one year prior to case/control selection was determined from Medicare claim files. Women with more than one SARD were considered to have an overlap syndrome (OLS, n=318). Associations with breast cancer (overall and by the tumor estrogen-receptor expression) were estimated by unconditional logistic regression adjusting for socio-demographic and health utilization variables. Results. Elderly women with RA were less likely to develop breast cancer (OR=0.87, 95%CI=0.82-0.93). The risk reduction of breast cancer associated with RA was statistically significant for estrogen receptor-positive (ER+) but not estrogen receptor-negative (ER-) tumors. In women with SLE, the risk of having ER-, but not ER+ (OR=0.49, 95% CI= 0.26-0.93 and 1.1, 0.86-1.4, respectively) breast cancers, was lower than the risk in women without any SARDs. Conclusion. The association between SARDs and breast cancer varies by SARD entities. SARDs affect breast cancer risk differentially according to the estrogen receptor status of the tumor. These findings support the hypothesis that the immune system alters breast cancer risk.