Further to zbnet's response, the 1221 event will give the worst-case estimate of available space. (i.e. the least you will regain after a defrag). Another option is to run eseutil with the /ms switch. this gives the best-case estimate of recoverable size. the actual value will fall between the two.
Dont forget that the edb file need white space to function. This is usually around 10 times the daily activity of the database (measured using transaction logs). Info for this can be found in Q195914. If the figure you come up with isn't that much, then dont bother.
We managed to defrag a 128Gb database down to 7.8Gb last week. This was due to job losses (deleted mailboxes), a severe audit I did and setting mailbox limits on the priv.edb. Sounds as though we had exactly the same issue as you. The worst users being senior management.
We managed to get around it by saying to them that exchange cannot be recovered in the event of DR** - not in the timeframe they were wanting. We then gave them an estimate of the cost of a new one (with licences, hardware and the cost of a full time employee to look after the increased admin costs). We then told them of a 'free' way of doing it, provided they conform. Although this worked, it still took them 18 months to delete mail. Regular hit-lists of offenders helps (they are usually at the top).
Good luck!
**I'm sure I read somewhere that Exchange recovery cannot be guarented for 5.5 if the edb file is over 60Gb.