Pre-diagnostic circulating carotenoids and lethal breast cancer risk: a novel analytic approach applied to a nested case-control study

Background Identifying pre-diagnostic modifiable factors that are associated with lethal breast cancer risk may provide prevention strategies for fatal disease. Objective We sought to examine whether pre-diagnostic circulating carotenoids and retinol concentrations are associated with a lower risk of lethal breast cancer using multi-state modeling. Methods This is a nested case-control study including 3,516 controls and 2,583 incident invasive breast cancer cases, among whom 190 developed lethal breast cancer (experienced recurrence or died due to breast cancer) over three 10-year intervals. Pre-diagnostic plasma carotenoids (α-carotene, β-carotene, β-cryptoxanthin, lycopene, and lutein/zeaxanthin) and retinol were assayed using high-performance liquid chromatography; total carotenoids were calculated as the sum of individual carotenoids. Based on a multi-state model that combined separate estimates for breast cancer incidence and recurrence/mortality due to breast cancer among invasive cases, we examined the association between pre-diagnostic circulating carotenoids and retinol and the probability of transitioning from disease-free to invasive breast cancer to breast cancer recurrence/death over 30 years. Results Compared to the lowest quintile, females in the highest quintile of β-carotene had a 41% lower risk of lethal breast cancer (β-carotene HR=0.59, 95% CI: (0.41, 0.83), p trend <0.01); for β-cryptoxanthin, HR was 0.68 (95% CI: 0.47, 0.97); for total carotenoids, risk of lethal breast cancer was 48% lower (HR=0.52, 95% CI: (0.36, 0.74), p trend <0.01), which for β-carotene and total carotenoids captured the combined protective effects of incidence and recurrence/mortality, while for β-cryptoxanthin, reflected protective effects of recurrence/mortality but not for incidence. For β-carotene, stronger inverse associations were observed for postmenopausal females (p heterogeneity <0.01), and marginally stronger for ever smokers (p heterogeneity =0.11), and for females with BMI<25kg/m 2 (p heterogeneity =0.10). Other plasma carotenoids or retinol were not associated with reduced lethal breast cancer risk. Conclusion Higher pre-diagnostic β-carotene, β-cryptoxanthin, and total carotenoid concentrations may be associated with a lower likelihood of progression from a cancer-free state to fatal breast cancer.