Try updating to Snap Manager for Exchange 6.0.4, it has some new bug fixes especially 518953:
518953 Title: When a new Storage Group/Database is added to the existing SnapManager for
Exchange (SME) backup management group, the SME backup retention is executed
based on the "least amount of backups found" for any Storage Group/Database
Description: For a backup job that contains two or more Storage Groups SG (or in
Exchange 2010, Mailbox Databases) that share a common LUN for Logs and SnapInfo,
SnapManager for Exchange (SME)deletes the oldest number of database snapshots
from each Storage Group LUN based on the "least number of backups found"
algorithm for any of the StorageGroup/Database in the backup group.
In a scenario when an Exchange Administrator adds a new SG/Database to an existing
Backup Job, and the new SG/Database shares the Log/SnapInfo LUN with the existing
SG/Databases, the number of backups for each will become asymmetric and not even.
As a result of current SME deletion algorithm system may end up in a temporary
situation that snapshots are not being deleted for a time period (despite the retention
policy), until the SG/Database with a least amount of backups exceeds the backup
group retention number.