OBJECTIVE: Non-steroidal anti-inflammatory drugs (NSAIDs) have been consistently associated with reduced colon cancer risk. Associations with other cancer sites, particularly breast cancer, have been less consistent. This study's objective was to examine the relationship between use of non-prescription and prescription NSAIDs and breast cancer and carcinoma in situ of the breast. METHODS: We analyzed data from a population-based, case-control study conducted in North Carolina between 1996 and 2000 to examine the association between NSAID use and breast cancer. Logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CI) separately for invasive breast cancer and carcinoma in situ of the breast. FINDINGS: NSAID use was inversely associated with invasive breast cancer (multivariable adjusted OR = 0.4; 95% CI: 0.3-0.6). Inverse associations were apparent for both prescription and non-prescription medications, and for occasional as well as regular use of NSAIDs. Inverse associations also were observed with carcinoma in situ, however the effects were somewhat weaker. INTERPRETATION: Our study adds to the accumulating, but somewhat inconsistent evidence that NSAIDs may reduce the risk of breast cancer. Future studies should work to define the most efficacious agents and determine whether genetic, lifestyle, or other characteristics influence their anti-carcinogenic properties.