reported. Cognitive reserve could have a role by modulating this relationship. The aim was to study the association between WML and the risk of mild cognitive impairment (MCI)/dementia, and to evaluate whether premorbid Intellectual Quotient (IQ), a marker of cognitive reserve, modulates this association. Methods: 547 individuals, without MCI/dementia, from a cohort of community-dwelling persons aged between 65 and 80 years (the ESPRIT study), received structural cerebralMagnetic Resonance Imaging (MRI) at baseline. WML volumes were assessed by a semi-automatic method on T2-weighted MRI, and analyzed as a categorical variable (divided into tertiles). Diagnosis of MCI was made according to revised algorithm (MCI-R) criteria and dementia cases were diagnosed by a neurologist and validated by an independent committee. These diagnostics were performed at baseline, 2, 4 and 7 years of follow-up. Premorbid IQwas assessed by the National Adult Test Reading (NART). The sample was dichotomized into low and high IQ subgroups according to the median score. To study the association between WML and risk of MCI/dementia, Cox proportional hazards models were performed. Results: 143 incident cases of dementia (n1⁄414, 2.6%) and MCI (n1⁄4129, 23.6%) occurred over the 7 years of follow-up. After adjustment for potential confounders (age, gender, Apolipoprotein E4 genotype, depressive symptomatology, vascular pathology, hypertension, education level, premorbid IQ, and total brain volume), participants with severe WML (last tertile) showed a significant increased risk to develop dementia and MCI over 7 years (Hazard Ratio (HR) 1⁄4 1.69, 95% CI:1.10-2.58 p1⁄4 0.02). After stratification on premorbid IQ status, the increasing risk of MCI/dementia related to WML severity remained significant in the low IQ group (HR1⁄4 2.02, 95% CI, 1.13-3.62, p1⁄40.02), while the association was not significant in the high IQ group (HR1⁄4 1.45; 95% CI, 0.76-2.76, p1⁄40.26). Conclusions: These results highlight the impact of WML on the risk of MCI/dementia in healthy elderly people and support the cognitive reserve hypothesis in demonstrating that subjects with high pre-morbid IQ are more likely to withstand deleterious effect of severe WML.