Abstract

INTRODUCTION: Some associations of high total cholesterol with dementia risk diminish as the outcome age-age at cognitive assessment-increases. METHODS: The Framingham Heart Study provided 1897 participants with intact cognition at entry. Cox regression analysis for incident marked cognitive decline included "time-dependent" coefficients, with associations between total cholesterol and covariates changing by outcome age. Decline within age categories of 75-84 and 85-94 years was also examined. RESULTS: Significant associations of rising total cholesterol linear slope, low entry age, low education, and statin nonuse with risk diminished significantly by outcome age. At 85-94 years, falling linear slope was significant. DISCUSSION: The protected survival model posits a minority subpopulation with protection against mortality and cognitive decline associated with total cholesterol risk factors. It predicts the observed diminished or reversed cholesterol associations with increasing age. Protection is particularly likely for successful cognitive aging-intact cognition at very old age-despite increased risk from cholesterol.