I actually do. Over the years I have seen a number of cases of metformin induced lactic acidosis. Although it is very hard to predict who will have it. I would like to be on the safe side. 

I see no harm in holding metformin prior a contrast study if there is enough lead time. That being said - not sure of any data if this strategy reduces the incidence of metformin-related metabolic acidosis.

Indeed Metformin she not be given to any CKD 3b patient given the high risk of AKI in these patients that may lead to imminent metformin-associated lactic acidosis with poor to fatal outcomes. See this NEJM case record: Kalantar-Zadeh et al., PMID 23841704.