Keep in mind that it is not likely that the 'shape' of the resulting ripple voltage is also going to be different with a multi-stage filter compared to a single.

It is my opinion that keeping the DCR low is a good idea. It isn't possible to do it 'all else kept the same' so it shouldn't follow that PS performance is tied to a the value of a single parameter. Either way, examine the performance of the filter as component values change around a reasonable centre, or preliminary value.
cheers,
Douglas