ObjectivesThe aim of this study was to evaluate the influence of an approaching cancer death on end-of-life aspirin use, a frequently prescribed medication for cardiovascular disease prevention.MethodsThis study was conducted using linked cancer registry and prescribing data. Breast (n=1151) and colorectal (n=1859) cancer decedents were matched to cancer survivors and the probability of either initiating aspirin, or continuing established aspirin use, was estimated in consecutive periods over the 5 years approaching a cancer-specific death (decedents) or matched index date (survivors).ResultsUsing the linked data sets, we identified patients who died of their cancer (decedents) between 1 January 2001 and 31 December 2009. In the 5 years prior to death, we compared (1) the probability of initiating aspirin use for the first time, and (2) the probability of continuing aspirin use. In comparison to matched cancer survivors, an approaching cancer death was not associated with a reduction in aspirin initiation by breast or colorectal cancer decedents. However, the probability of continuing established aspirin use declined considerably in the 24 months approaching death and at the time of a death was significantly lower for breast (risk difference (RD) −0.26, 95% CI −0.33 to −0.20) and colorectal (RD −0.38, 95% CI −0.46 to −0.30) cancer decedents versus matched survivors.ConclusionA significant proportion of patients discontinue their aspirin in the time approaching a breast or colorectal cancer-specific death. The safety and benefits of this are unclear and empirical data are needed to guide decisions about aspirin use in the end of life.