What is the difference in the implementation of a monad in a purely functional language with respect in imperative language?

For a long time, the use of these structures Monad has been limited to a very small circle of languages, many of which are purely functional (mainly due to problems related to the management of information operations).
Recently, these structures have also been adapted to imperative languages

My question is this: we are talking about the same model, Monad, but what is the difference in the implementation of a monad in a purely functional language compared to the same model but implemented in imperative programming?

What change? Of course, this does not change the idea, but if we are talking about the same implementation of the design model but in 2 different programming paradigms, something must be different, probably the same information will come back, but we use an identical idea but in 2 contexts of different paradigms the effect can not be the same, otherwise it would not have seemed to us to apply the monad for imperative language