Ten studies were included (n>236,655): four cohort studies (n>185,700) and six case-control studies (n=50,885). There was no evidence of publication bias.
Aspirin exposure was associated with a statistically significant reduction in breast cancer risk (RR using random-effects model 0.75, 95% CI 0.64 to 0.88). Significant heterogeneity was found (p<0.01).
Meta-regression found no influence on risk of either study type or population. Case-control studies found a slightly stronger association between aspirin and breast cancer than cohort studies, but the difference was not statistically significant in the meta-regression. There was no evidence that the definition of non-exposure influenced the estimated risk of breast cancer.
Both methods of scoring the upper open-ended aspirin intake showed that a higher frequency of aspirin use was associated with a significant reduction in the risk of breast cancer (p=0.002 and p=0.0008). Longer duration of aspirin use was associated with a more protective effect than shorter duration (p=0.06 and p=0.04). Pill years showed a significant and slightly protective effect on the risk of breast cancer (p=0.002 for one year and p=0.0008 for two years).