Dear Alfonso,

I assumed that regression model will be ill posed, if we use correlated or non-orthogonal variables in the model.
Since we are not interested in the weights of confoundings, you are right redundancy won't be a problem.

Thank you so much.
Regards
Ali