Late-life dementia results from non-modifiable risk factors such as age and genetics, modulated by deleterious and protective environmental factors among which nutrition may play a major role. This paper highlights five major recent contributions of the French Three-City (3C) and PAQUID epidemiological studies to Alzheimer's disease (AD) knowledge, targeting genetic and dietary risk factors, and the impact of cognitive decline in daily living. The 3C study contributed to a large genome-wide association study to identify new genetic risk factors for AD. In addition to apolipoprotein E (APOE), two loci gave replicated evidence of association: one within CLU, encoding clusterin or apolipoprotein J, and the other within CR1, encoding the complement component receptor 1. Although the attributable fraction of risk for these polymorphisms is moderate, genetic studies provide significant insights into the molecular bases of AD. Regarding dietary data, findings from 3C suggest that healthy diets associating sources of both omega 3 fatty acids (fish) and antioxidants (fruits and vegetables) such as the Mediterranean diet, and caffeine could be associated with decreased risk for AD. However, the protective effect of omega3 fatty acids might be limited to APOE4 non-carriers. Future research should focus on gene-nutrient interactions. Regarding the functional impact of prodromal AD, the PAQUID study showed that taking into account mild functional limitations considerably increases the predictive value of neuropsychological tests for conversion to dementia. Research should focus on sensitive instruments to capture early functional decline to improve the identification of elderly patients at high risk of conversion to dementia.