No downvote from me, but couldn't it be argued that the more cycling is taken up, the likelihood is that fewer people would be driving so there'd naturally be less exposure? So then it would be true that the more cycling is taken up, the safer it becomes -- because there are fewer cars driving about.

As you mentioned, there's less vehicle exposure in countries where there are low cycle death rate -- lower car usage is presumably a big factor here, though well-designed road sharing and separate cycle paths etc. would presumably also help a lot. (Of course, people can be killed cycling in other ways than being hit by a car, but I'm guessing that's a small minority of cases -- would be interesting to see figures though.)