This is an implementation detail, and doesn't affect the API presented to the user. For example, consider the clj-http library that's backed by the Apache http Java library internally https://github.com/dakrone/clj-http . As a user of clj-http, I don't know or care about the fact that it's a wrapper around OOP code.

Furthermore, if you wanted to, you could reimplement the internals in Clojure and the users of the library wouldn't be affected.

Being able to leverage existing mature libraries is a huge advantage, and the fact that they can be used via idiomatic functional APIs provides the best of both worlds experience.