Hmmm...
I remember doing some (non-professional) research on marital status and life outcomes, and I recall there being a correlation between marital status and life expectancy -- marriage led to longer lifespans.
I could be mistaken.
But if I'm not, could it be that the unmarried die earlier, and thus are less likely to have acquired dementia of any type?
The associations for divorced and never married remained significant accounting for demographic, behavioral, clinical, genetic, referral source, participation, and diagnostic factors. The associations were slightly stronger among professional referrals, males, and relatively younger participants.
(Not sure what "demographic" covers...)
Males also tend to live shorter lifespans, especially when unmarried, IIRC...