Then of course is the possibility of making N-formyl-(MD)amphetamine using ammonium formate/(MD)P2P, and then reduce the amide to (MD)MA with NaBH4 plus a metal salt (ZrCl4, CoCl2, NiCl2, LiCl etc).
That way you can arrive at (MD)MA from (MD)P2P without the need for any methylamine, and the yields should be similar to that from the Al/Hg reductive amination.