This is a regulatory rather than a scientific issue.
From a scientific perspective, the answer depends on how well the system suitability requirements were written. Since your are using "L" designations for columns, I assume this is a USP method. Many USP methods are old, and system suitability was an afterthought; in essence a last-minute check for gross problems. In other words, you can meet system suitability and still turn out erroneous results. On newer methods, system suitability is more likely to be meaningfully tight.
From a regulatory perspective (which is more important here), the answer depends on your SOPs.
- - If your SOPs state that you will follow the method, then you're out of compliance, period. If I were an auditor, this would raise a gigantic warning flag, and I would go back and look very hard at all of your record-keeping, training, qualification, etc. records.
- If you have an SOP that says you can use an equivalent column as defined by USP, then you're still out of compliance, although you can make a case for equivalence on the basis of system suit, as mentioned above (the whole issue of column equivalence is the topic of serious discussion within USP at the moment).
- If you have an SOP that states that meeting system suitability implies the column is suitable (certainly a justifiable statement) then you're OK.
- If you don't have an SOP on column substitution, then write one!