There’s no perfect answer for this question and it depends on individual.
You had used address A (one of the outputs) in a transaction 1 where inputs are from X.
You used the address A (one of the outputs) again in a transaction 2 where inputs are from Exchange Y.
IF someone is trying to analyze transaction 1 to know the owner of address A, he can look at transaction 2 to assume owner has an account on Exchange Y. Most of the withdrawal transactions from exchanges are done in batches with lots of inputs and outputs involved so it won’t be easier to know more information about the owner except if one of the below tries to investigate:
Exchange, chain analysis companies, government agencies (in case of centralized exchanges) etc.
You have two options in my opinion:
- Ignore and avoid reuse in future