I have successfully used HFBA in LC-MSn at concentrations as high as 20 mM. The MS pretty much becomes a dedicated system, and don't even think about doing negative ion analysis, but it does work. I had originally tried a lower concentration (5 mM), but analyte retention times weren't reproducible until I bumped up the HFBA. The final method (for multiple aminoglycosides in meat) can be found on USDA FSIS's website:
http://www.fsis.usda.gov/OPHS/clg/index.htm (#16)
If I had to do aminoglycosides by LC-MS again, I'd give HILIC a serious try. You have to use really high NH4OAc (0.2 M!) to get it to work (so far I've tried Waters column and have ZIC column, but haven't tested that one, yet). 0.2 M seems outrageous on an LC-MS, but you start out at 100% ACN, so the buffer is fairly diluted. Much easier to wash the NH4OAc out of the LC-MS system than HFBA (I've heard you actually have to replace LC pump vacuum degasser membrane to really get rid of the HFBA). Bonus was that I had the best response for aminoglycosides that I've ever seen on our systems.