Allergies run the gamut of being deadly, in the case of some bee sting cases, to sickening and may be related to autoimmune diseases:
https://www.nih.gov/news-events/news-releases/nih-scientists-find-link-between-allergic-autoimmune-diseases-mouse-study Overall, I suspect it's a bit like anything else, with allergen susceptibility genes making their way in the population especially as conditions occur(ed) that minimize their effect on mortality.
You noted: "Something so deadly the smell or particulate requires an epi-pen for survival? No genetic reproduction there."
And that's just it,....when epi-pens didn't exist, there was a certain amount of mortality associated with a severe allergic reaction. If situations change where these reactions are no longer detrimental to reproduction, then there will be no selective action removing that susceptibility from the gene pool.
"I'm not highly fond of the adaption method- I think that just makes your immune system readier, but still battling."
I'm not sure of the basis for the success of the adaptation method, but have heard from some for whom it worked. My on-again, off-again hay/pollen/dust allergies have declined over the years, but I think an early mild cat allergy has perhaps gotten worse. So in some cases they can be out-grown, but probably not in all.