You give no parameters such as water volume of the system, nor bio-load. 55 gallon barrels using this methodology (with a strict implementation criteria which does not include a 'fuge', e.g. rock, macro, critters) is used on systems in the 2000 gallon range....the barrel filled to the point, with oolitic aragonite sand, that only a few inches of water above the sand remains. It is staggering how quickly this method will reduce the nitrates in volumes this size.
Beyond that there really isn't an answer that would be based in any sort of quantitative or qualitative anlysis, to determine what is too deep, or not deep enough.
The 'model' for this is much more modest, and uses a 5 gallon bucket with 60lbs of oolitic sand, and that fills it up leaving only a few inches above the sand, over which water flows. This will serve around 120 gallons. So to do this right, (in 55 gallon drum) you are looking at ~ 600 lbs of sand...
A lot of folks believe that combining a sand bed with a 'fuge' is somehow a good idea. However, adding rock, macro, and the associated life that comes with it, changes the needed methodology for a dsb, from a simple, maintanence free, unlit pile of sand, to the need for a method that can deal with the inevitable particulate matter issues that will turn the dsb into a mess: in need of care and feeding every bit as intensive as it would be in a DT. Whereas a 'bucket dsb,' as it is refered to, will eliminate entirely the need or desire, even in the most remote sense, for a 'fuge,' unless you are wanting to keep seahorses or similar, and you would not be doing that in a 55 gallon drum, or at least I hope not.
So, I would have to say that if you don't have a 2000 gallon system, then a 55 gallon drum is a huge waste, and at the other end of the scale, a little fuge (whatever that means) and sand bed in a sump is next to useless.