I think we may have a terminology problem here. Are you actually validating a method or running an existing validated method?
First of all, I don't think you can "fail" a system suitability test during validation, because defining system suitability criteria is part of the validation process.
Second, what do you know about the history of that column? Has it been used for other things? Has it been properly stored? You cannot properly validate a method using a column of dubious provenance; there is no guarantee that previous use has not changed its properties.
Third, an important part of validation is determining the robustness of the method, including robustness with respect to column chemistry. That means that you should be running with at least two (preferably three) columns from at least two identifiably different batches or lots.
If you are simply trying to verify or run the method (rather than fully validating it), yes it is possible that the column has been damaged either through abuse or improper storage. That said, it is possible that the problem is something physical such as extra-column volume from large-bore tubing or an improperly seated tube end in a connecting fitting. If this were my problem, I would check all the connections and then re-test with the manufacturer's test procedure on the CoA that came with the column. If it fails, get a new column. Columns are cheap, your time is valuable! It's very easy to spend more time than the column is worth.
