INTRODUCTION: We tested the hypothesis that low plasma complement C3 is observationally and genetically associated with high risk of Alzheimer's disease (AD).

METHODS: We studied 95,442 individuals enrolled in the Copenhagen General Population Study. In genetic analyses, we further included 8367 individuals from the Copenhagen City Heart Study. In the two studies, 1189 and 35 developed AD during median 8 years follow-up.

RESULTS: The multifactorially adjusted hazard ratio for risk of AD for a one standard deviation lower levels of complement C3 was 1.11 (95% confidence interval: 1.04-1.19) in all individuals and 1.66 (1.33-2.07) in APOE ε44 carriers. In Mendelian randomization, the corresponding genetic estimates were 1.66 (1.05-2.63) overall and 1.99 (0.52-7.65) in APOE ε44 carriers.

DISCUSSION: Low baseline levels of complement C3 were associated with high risk of AD. The risk was amplified in APOE ε44 highly susceptible individuals, and these findings were substantiated by a Mendelian randomization approach, potentially implying causality. Based on these findings, we present and thoroughly discuss an updated Alzheimer hypothesis incorporating low complement C3 levels.