Two reasons why this hypothesis may not be true:

1) Universal failure of plaque-reduction drug trials.

2) Numerous post-mortem counter-examples: dementia w/o plaques and plaques without dementia.