Purpose: Cross-sectional studies show that statins, used in cardiovascular disease prevention, are often discontinued approaching death. Studies investigating associations between statin exposure and cancer outcomes, not accounting for these exposure changes, are prone to reverse causation bias. The aim of this study was to describe longitudinally the changes in statin initiation and continuation prior to death in patients with breast or colorectal cancer, thus establishing an appropriate exposure lag time.
Methods: This study was carried out using linked cancer registry and prescribing data. We identified patients who died of their cancer (cases) and cancer survivors were used as controls. The probability of initiating or continuing statin use was estimated up to 5 years prior to death (or index date). Conditional binomial models were used to estimate relative risks and risk differences for associations between approaching cancer death and statin use.
Results: Compared to controls, the probability of continued statin use in breast cancer cases was significantly lower 3 months prior to death (RR 0.86 95% CI 0.79, 0.94). Similarly, in colorectal cancer cases, the probability of continued statin use was significantly lower 3 months prior to colorectal cancer death (RR 0.77 95% CI 0.68, 0.88).
Conclusion: A significant proportion of patients will cease statin treatment in the months prior to a colorectal or breast cancer death.
Keywords: Breast cancer; Colorectal cancer; End-of-life; Epidemiology; Statins.