Lipophilic Statins in Subjects with Early Mild Cognitive Impairment: Associations with Conversion to Dementia and Decline in Posterior Cingulate Brain Metabolism in a Long-term Prospective Longitudinal Multi-Center Study

102 Objectives: Studies of the relationship between use of statins (HMG-CoA reductase inhibitors) and subsequent cognitive performance have been variously reported to demonstrate beneficial, harmful, or no significant effects. We aimed to help clarify the relationship between statin use and subjects’ long-term cognitive trajectory in subpopulations prospectively and longitudinally examined, as stratified by 1) neuropsychological status at baseline, 2) relatively high vs. low serum cholesterol levels at baseline, 3) statin use vs. non-use and, among users, type of statin used. The present analysis focuses upon outcomes of subjects with early mild cognitive impairment (eMCI), comparing users of statins with known moderate (atorvastatin) or high (simvastatin) lipophilicity and blood-brain barrier penetrance (LS), to non-users (nonS), or users of other statins (OS). Methods: Subjects were drawn from a consecutive series enrolled in the Alzheimer’s Disease Neuroimaging Initiative at over 50 North American sites. Of a total of 392 eMCI subjects, 303 had cholesterol levels available at baseline. Subjects were then grouped into those above (n=103) or below (n=200) the median cholesterol level of the nonS subjects (206 mg/dl). The significance of differential dementia conversion rates of eMCI subjects was assessed by Chi-Squared tests. Statistical parametric mapping of FDG PET scan data was used to perform paired t-test analyses to identify any regions of declining cerebral metabolism within each statin group. Results: While serum cholesterol levels at baseline ranged widely (101-358 mg/dl), among all eMCI subjects the average baseline cholesterol levels did not significantly differ between those who did convert to dementia within 96 months vs. those who did not; after excluding statin users with less than 96 months of use, the below-median cholesterol group consisted of 157 subjects (67 nonS, 72 LS, 18 OS), and within this group, average serum levels again did not differ between those who did (166 mg/dl) and did not (171 mg/dl) convert to dementia. There was, however, a significant difference observed in conversion rates within this group according to statin use: among LS, 24% converted to dementia, vs. only 10% of nonS (p=0.04) in the ensuing 96 months, while conversion rate of OS did not significantly differ from nonS subjects (11%, p=0.94). Moreover, posterior cingulate metabolic decline was identified among LS users, (p<0.0005, highly significant after statistical correction for multiple comparisons), while no significant decline occurred among OS and nonS subjects. Finally, in the above-median cholesterol stratum, the difference in conversion rates of statin users and non-users was not significant (p=0.72). Conclusions: Among subjects with early mild cognitive impairment and low to moderate serum cholesterol levels at baseline, lipophilic statin use was associated with more than double the risk of converting to dementia over eight years of follow-up compared with statin non-use, and with highly significant decline in metabolism of posterior cingulate cortex -- the region of the brain known to decline the most significantly in the earliest stages of Alzheimer’s disease. In contrast, no such clinical or metabolic decline was found for users of other statins, nor statin users having higher baseline serum cholesterol levels.