Another great way to deal with this is using the custom visual "Chicklet Slicers" by Microsoft. I had this slicer issue - mine included not only blanks, but also other records represented in the slicer that had a null or zero value for the measure I was hoping to filter them by. Using the chicklet slicer, the report not only looks better, but easily filtered out the records in the slicer that did not have a value by dragging the measure I was using into the data field of the slicer visual.
I went to investigate and found that the company that developed this (in many ways rather excellent) custom visual still has "ragged hierarchy" (which means that "empties" are not displaid) on their development list. I hope they get to it soon, or that MS makes this visual standard with a few enhancements.