Cholesterol and cognitive performance in normal controls and the influence of elective statin use after conversion to mild cognitive impairment: results in a clinical trial cohort

Abstract

Background: We reported a significant 67% reduction in the hazard risk of incident Alzheimer's disease (AD) with elective statin use in the AD Anti-inflammatory Prevention Trial (ADAPT), without a reduction in risk of incident mild cognitive impairment (MCI).

Objective: To assess if cholesterol levels are associated with cognitive performance and determine if statin use alters cognitive performance after onset of MCI.

Design: Fractionated cholesterol levels, neurological and cognitive status were evaluated annually. Comparisons of non-LLA (lipid-lowering agent) users or statin-LLA users were performed blind to the ADAPT medication randomization. Pearson's correlations were validated using a time-dependent linear mixed model.

Results: The MMSE performance significantly declined over time in non-LLA users, and, after adjusting for this, a significant positive correlation between MMSE and HDL was identified (p = 0.0002). A negative correlation between total and LDL cholesterol, and immediate and delayed recall of the Rivermead paragraph was significant (total cholesterol, p < 0.003; LDL, p < 0.02). Pilot data suggest a positive signal on delayed recall of both the Hopkins word list and Rivermead paragraph with deterioration in the non-LLA users and improvement in the statin users after conversion to MCI.

Conclusion: Cholesterol levels may be associated with differential performance on the MMSE and measures of learning or memory. The trend for improved delayed recall in statin users with MCI compared to non-LLA users with MCI may have contributed to the reduced hazards risk of incident AD without reducing the risk of MCI.

Fig. 1

Plot of estimated performance on delayed recall of the Hopkins word list among individuals with MCI taking statins (dashed line) and those not taking an LLA (solid line) versus time (0–30 months). p value for slope of the statin line is 0.083; p value for slope of the non-LLA line is 0.24; p value for difference in slope is 0.039.

Fig. 2

Plot of estimated performance on delayed recall of the Rivermead paragraph among individuals with MCI taking statins (dashed line) and those not taking an LLA (solid line) versus time (0–30 months). p value for slope of the statin line is 0.23; p value for slope of the non-LLA line is 0.23; p value for difference in slope is 0.096.