Alzheimer's in particular does not reduce lifespan all that much; rather, it increases the resources that children and grandchildren need to use to care for the parent. So selection against Alzheimer's would presumably (we have very little evidence for any of this) be even stronger than selection against traits that, say, killed at 70.
Is actually caring for an Alzheimer's stricken relative a new phenomenon? Could it be that a hundred years ago, a chronically ill, continually deteriorating relative was just left in a room to waste away?
Professional Alzheimer's care is new;, without good professional work you're going to get sores and muscle atrophy and all kinds of other complications that can be fatal at old age. We've gotten very good at dealing with those, but not at actually curing the underlying disease. (Speaking from family experience.)
Maybe even more importantly, though, 100 years ago other, unrelated diseases of old age were likely to get you well before Alzheimer's really got bad.
