I'll answer my own question... again.
aoctool can't take more than one covariate and there's no way round it. However, there is a file on the file exchange called MANCOVAN which can do all sorts of things - including an ANCOVA with more than one covariate.
That function does not output the stats array required by multcompare though, so really it's a bit of a dead end. Matlab really sucks sometimes.