Yes you are right, I forgot the tags are not part of the cell once placed in a drawing.
I've had no luck at all coming to a solution writing tags back to the DGN via the variant array so this morning I decided to perform a little test by changing the scan criteria back to look for msdElementTypeTag directly instead of msdElementTypeCellHeader and searching the host element for any tags. I'm glad to report that I've been able to read and write updated tagdefinition values to dgn files successfully.
I've had no luck at all coming to a solution writing tags back to the DGN via the variant array so this morning I decided to perform a little test by changing the scan criteria back to look for msdElementTypeTag directly instead of msdElementTypeCellHeader and searching the host element for any tags. I'm glad to report that I've been able to read and write updated tagdefinition values to dgn files successfully.