BACKGROUND: Clarifying the aetiology of dementia is of crucial importance in the management of patients as well as for research purposes but it is not always possible clinically. Therefore the identification of biological markers should complement clinical approaches. Telomere shortening is emerging as an important mechanism in vascular aging and the pathogenesis of hypertension and atherosclerosis. Thus, telomere length could be a potential candidate to accurately separate vascular from degenerative cognitive impairment. OBJECTIVES: To evaluate the usefulness of telomere length alone or combined with ApoE polymorphism in diagnosing mild cognitive impairment (MCI) and in differentiating Alzheimer's disease (AD) from vascular (VaD) and mixed dementia (MD). METHODS: Telomere length in peripheral blood lymphocytes was performed by flow cytometry in 439 patients (mean age, 85.1 years): 204 cognitively normal, 187 demented patients: 80 AD, 86 MD, and 21 with VaD; and 48 patients with MCI. Simple and multiple ordered logistic regressions were used to predict the risk of dementia from telomere length, ApoE polymorphism and age. RESULTS: ApoEepsilon4 was statistically associated with patients with dementia (p<0.001) compared to cognitively normal or MCI patients; but not with the aetiologies of dementia (AD, VaD and MD) (p=0.385). No significant differences in telomere length were found among patients with different aetiologies or severities of dementia. In the global model, the combination of telomere length and ApoE polymorphism did not confer a significantly higher dementia risk (OR=0.95, 95% CI=0.69-1.32; p=0.784) than APOEepsilon4 alone (OR=2.12, 95% CI=1.15-3.9; p=0.016). CONCLUSION: This longitudinal study in very old patients provided no evidence suggesting that telomere length alone could be used to distinguish between the different types of dementia or MCI, nor combined with the ApoE polymorphism.